February 3, 2016
UMass Lowell (in conjunction with 7News) is running a daily tracking poll. These are fun. In my youth, I worked for a market research company that did the dialing for a GOP-affiliated pollster. Near election day, you really did see the numbers move. The 20+ year Senate career of Barbara Boxer almost never happened. One last-minute semi-revelation about her opponent and a career was born.
An abridged guide to the polls of New Hampshire is coming later in the day, but we’ll pay extra attention to the tracker as a measure of shifts. We won’t know until they vote if UMass is any more accurate than any other pollster. I don’t actually care. As long as one day isn’t more polluted than the next, it serves the purpose.
This tracker is a three day weighted average. Each day they toss one day out and add in a new one. In this case, our data is from January 30 thru February 1. This makes it completely free of Iowa caucus result influence. On the 5th, we’ll see the numbers from 2/2-2/4 and have a full post-caucus look.
Our deep dive is on the GOP side. We’ll talk about the other contest separately. Tracking polls are great, but crosstabs from tracking polls are even better. This will let us see how the various voter groups move over the next week. For my first act, I’ll attempt to shock you with a few crazy-ass splits.
You know how Donald Trump does better with less educated voters? You know how John Kasich is positioned similarly to Jon Huntsman in 2012, and might expect to do better with more educated voters? Check this out:
High School or Less
Post Graduate Degree
Watch this going forward. If Trump gets over 50% among less educated voters, he’s going to hang on and then some. Kasich can’t finish a strong second getting 2% from a big voter group.
Side Note: UMass gives Trump higher overall numbers than average and Kasich lower. That shouldn’t impact what we’re seeing here. They aren’t making their inclusion decisions based on education level.
One of the most important questions is how locked in a voter is. We know Trump voters are more likely to claim they have their mind made up. We saw in Iowa that less committed Trump voters may have abandoned him. Using these numbers is a good way to determine a temporary floor for each candidate.
I’ll also include a measure of how much this changed from the previous day’s tracking info. You’ll notice a definite supporter isn’t permanently definite.
Trump 69% (-3)
Kasich 65% (even)
Cruz 58% (-7)
Rubio 57% (+14)
Bush 50% (+9)
Christie 49% (+15)
Paul 49% (+7)
Carson 27% (-7)
Again, this is pre-Iowa results. The changes are larger than they seem. It means switching out one of three days in the sample created an adjustment of this amount.
Rubio was making progress in New Hampshire ahead of his third place “victory.” This is a relatively high level of certainty for his voters, who are normally less committed. In Iowa, he apparently locked down all of his leaners, plus took some from Trump and/or others.
At first glance, Cruz looks like he was moving backwards. One other consideration. If a candidate is increasing overall support, they may temporarily drop in this measure. Here’s how:
Cruz was at 12% overall and 65% certainty. That’s 7.8% dedicated support.
The next reading gave him 14% overall and 58% certainty. That’s 8.1% dedicated support. He didn’t actually lose any committed supporters. He just gained leaners faster. A bounce from Iowa could move leaners to committed, add new people as leaners, or both.
Rubio improved in both measures, going from 3.4% dedicated support to 5.7% dedicated. As a comparison, Trump currently sits at 26.2%. Marco still has some work to do, at least according to UMass.
You can find better or worse numbers for any candidate in a different poll, but it gives you some idea of the fragility of the candidate positions. Rubio is arguably the leading establishment-friendly candidate in an establishment-friendly state, and was making progress in New Hampshire too ahead of the Iowa vote. He had less than 6% firm polling support in this survey.
Polls are great. Tracking polls are even better. They give us useful data. Stuff will change noticeably every 24 hours. Gotta love New Hampshire Week.