Last week, a poll commissioned by OPB and Fox 12 suggested very different results in the Oregon presidential primary than what actually happened when ballots were counted Tuesday.

The poll showed Hillary Clinton leading Oregon's Democratic Primary among 48 percent of those surveyed. Last night, Vermont Senator Bernie Sanders won the state with 56 percent of the vote.


John Horvick of the polling firm, DHM Research, joined Think Out Loud Wednesday to discuss the poll and the actual results with guest host April Baer.

Q&A with John Horvick

April Baer: Rolling Stone’s Tim Dickinson — an Oregonian — described the presidential polling numbers as “extravagantly wrong.” The numbers showed Hillary Clinton would win the state by seven to 15 points. Sanders is currently winning by 12 percentage points. What happened?

John Horvick: You know, you gotta have some humility. We missed this one.


There are a couple things that we are certainly going to be looking at. First, this is the first election since Oregon has changed to motor voter, and we had a massive change in voter registration leading up to the election, and certainly we're going to want to look back and see what we can learn about our modeling in the future and how things shook out in this campaign, or in this particular election, with regards to that.

Related: Bernie Sanders Wins Oregon, But What Happens In November?

The other thing is that we saw — not just in our survey but in surveys and election results throughout the campaign so far — is that there is this very, very sharp difference between younger voters and older voters in their support for Sanders versus Clinton, and it's always difficult to predict with a poll who's going to show up to vote.

And we knew that, that's why we modeled it a couple different ways, that younger voters, the more they showed up, the more it was going to favor Sanders.

We're certainly going to be looking back at that too, just to try to get a better understanding of those two factors.

AB: Going to some of the Portland metro measures, you found that gas tax was leading 55 to 38, when the difference was two percentage points. What happened there?

JH: That's one where I feel much better about. So we were a few points there on the yes, and we tend to see, when things get closer to election day, things tighten up, so that followed a very typical pattern on that one. And actually, beyond the Democratic Primary race, the poll results that we came up with were in line with what the election results were, it was that one Democratic race that's the one that's going to be gnawing at us.

... Primary elections are particularly difficult because you’re trying to imagine through some past history what the turnouts going to be. And in the city races, there wasn’t the energy and interest that we’ve seen in past mayoral elections or past city commissioner elections so we had much higher levels of “don’t know” than is typical. But generally speaking, for the city of Portland surveys and the other races in our state surveys, we did quite well.