Allow me to apologize at the outset. Sometimes I get excited about data and about presenting that data, and sometimes that causes me to go a little overboard.
I went a little overboard.
With the Iowa caucuses out of the way, we turn our attention to New Hampshire, which is a much different state on both the Democratic and Republican sides of the presidential fight, but for very different reasons. To illustrate that, I took five sets of data and combined them: Demographic representation for each party in the last competitive Iowa caucuses (2008 for the Democrats and 2012 for the Republicans), representation in the last New Hampshire primary from those same years, the winner of each demographic in Iowa this year, and the margin by which they won, and how representation changed among the demographic groups in the Iowa voting. That's a lot, and the graphs are complex as a result. (As I said, I went a bit overboard.)
But I think they're also illuminating.
On Monday, Ted Cruz won Iowa, leveraging strong support from evangelicals and very conservative voters to upset Donald Trump, whose base of support was a smaller percentage of the total turnout.
New Hampshire is very different.
The colored circles are the main thing here. They're colored by who won the demographic in Iowa this year, and scaled to the margin of victory. (A big red circle means Cruz won that demographic by a lot; a little yellow won means Trump won by a little.) They're positioned according to how much of the vote they represented in Iowa and New Hampshire in 2012 -- but the representation in Iowa this year is indicated with a small gray circle. An arrow to the left means that the density of that demographic in Iowa fell this year; an arrow to the right means it increased. (The bubbles are positioned according to 2012, because that's a more even comparison with New Hampshire in the same year.) Then there's that dashed diagonal line. Any dot above that line indicates that representation from that group was higher in New Hampshire in 2012. A dot below the line? That demographic was more dense in Iowa.
What's immediately obvious is that two groups won by Donald Trump -- non-evangelical voters and moderates -- were much more of the electorate in New Hampshire in 2012 than they were in Iowa. Nearly four-fifths of the turnout in New Hampshire in 2012 wasn't evangelical, versus less than half in Iowa. But the density of that group fell in Iowa this year, as the density of the evangelical vote increased.
You can see that in the two large red circles at the bottom of the chart. That's how Cruz won: The size of those circles. But in 2012, those groups were much more common in Iowa than in New Hampshire, which is likely why Cruz is not polling anywhere close to Trump in that state.
It's worth noting that green circle near the middle is for Marco Rubio. He did well with "somewhat conservative" voters, who were more of the electorate in Iowa this year than in 2012. That's Rubio's strength. Conservatives flock to Cruz and moderates to Trump. In the middle, Rubio does well.
Notice that the density of those with only a high school education is very low. That's Trump's other strong demographic, but they aren't much of the vote, and they were more of the vote in Iowa in 2012 than in New Hampshire.
Notice, too, Trump's failure on turnout. The density of the vote of all of the groups that he won decreased from 2012 to 2016. Part of that was that Cruz flooded the zone and turned out his supporters. Part of it was that Trump's field program collapsed.
The result in Iowa for the Democrats was essentially a tie. Hillary Clinton emerged with slightly more delegate-equivalents than Bernie Sanders, but the distinction doesn't mean much. The two were pretty evenly split overall.
So how does it compare to New Hampshire?
The difference between Iowa and New Hampshire for the Democrats is much smaller. New Hampshire is a little wealthier, as you can see from the "$100,000 plus" income bubble above the diagonal line. Clinton does well with that group.
Notice that with one big exception -- moderates -- groups with which Clinton did well were usually more of the electorate in 2016 than in 2008. Older voters and wealthier voters were more of the electorate this time around than eight years ago.
Clinton did better with those calling themselves "somewhat liberal," positioned just above moderates. But Sanders, as we know, did much better with very liberal voters and young people -- the latter of whom didn't turn out as much for him as they did for Barack Obama.
Again, though, notice how close all of these bubbles are to the diagonal line. There is often not much difference in the electorate between the two states, unlike on the Republican side. So why is Sanders leading so much there? His campaign insists that it isn't geographic; you may draw your own conclusions.
There's a lot more to suss out from those graphs, by the way. Dig in. And if you think they're ridiculously over-stuffed and off-putting -- totally fair.
I went overboard.