Fractured GOP Field Is Pushing Trump to Nomination
In a night filled with surprising statistics, three pairs of numbers in particular stood out on Super Tuesday: 56, 56, and 47.
These are the percentages, respectively, of voters in Arkansas who said that they would be satisfied if Ted Cruz, Marco Rubio, and Donald Trump won the Republican nomination.
You read that correctly. A majority of voters in Arkansas were fine with either Rubio or Cruz as the nominee, while a majority of voters in Arkansas were not okay with Trump as the nominee (50 percent answered “no”). Yet Trump won the vote there by three percentage points, along with a near-majority of the state’s delegates.
We see this in other states. Consider Virginia. The numbers there are perhaps even more stark. Forty-three percent of Republicans said they would be satisfied if Cruz were the nominee, 59 percent said they would be satisfied if Rubio were the nominee, while 44 percent said they would be satisfied with Trump as the nominee. Yet Trump again won the state by about three points.
Or look at Texas: 68 percent were satisfied with the prospect of a Cruz nomination, 59 percent with the prospect of a Rubio nomination, and only 43 percent with the prospect of a Trump nomination. Yet Cruz ran 25 percentage points behind the share of voters who would be satisfied if he won the nomination, while Rubio ran 40 percent behind his respective number. Trump, on the other hand, ran only 18 percentage points behind the overall share of the electorate that said it would be satisfied with him winning the nomination, thus claiming a large number of delegates.
How is this possible? Part of it stems from the fact that Trump’s “suck the oxygen from the room” media strategy has prevented any particular candidate from breaking through: People may be “fine” with Rubio or Cruz, but they don’t necessarily know enough about them to feel passionately about either one.
But the main reason for this outcome is probably something I’ve been talking about for over a year, and which David Byler and I discussed Tuesday: The fractured Republican field is skewing the results. Tuesday night, Trump won seven contests, Cruz won three, while Rubio won one. Had it been a race between just those three candidates (or perhaps substituting Kasich for Rubio), it probably would have looked much different: Cruz winning four, Trump winning four, and Rubio/Kasich winning three.
What do I mean by this? Let’s look at the exit-poll data from several states. The first demographic I want to explore is education.
Now, these charts are a little counter-intuitive, and demand some explanation. These statistics represent the share of the vote won by candidates in a particular demographic, relative to their share in other demographics. In other words, if Ben Carson wins 6 percent of the vote among those with high school education, some college, college degrees, and postgrad degrees, he will receive a 25 percent in each group, showing that his support is evenly divided.
If, however, he received 6 percent of the vote among those with high school educations, but 1 percent in the remaining demographics, he’ll receive a 66 percent among high school-educated voters, and an 11 percent in the remaining fields. Think of it this way: If the latter numbers are true, then Carson is six times stronger among those with high school educations as he is among those with college educations. At the same time, he is equally strong among those with college educations as he is among those with, say, postgraduate educations. This method of calculation illustrates that.
What this shows us are the relative strengths and weaknesses of the various candidates across demographics. In other words, it can be useful to know whether Carson’s support is concentrated among high school students, or is evenly dispersed. Again, you can’t extrapolate directly from these numbers since, for example, there are relatively few voters with postgraduate degrees. This only gives a sense of how the candidates’ coalitions shape up.
I’ve also bolded and underlined the candidates’ strongest demographic showings in these charts to further highlight the patterns. I’ll emphasize that if you are looking for “hard” statistical proof of this, you can use the demographic calculator that David and I put in the piece linked to above. But I think this approach might be easier to understand.
So here is how the candidates’ coalitions were dispersed among educational groups.
There’s some variation here, especially among the Carson/Cruz demographics. But note that Rubio and Kasich pretty consistently show the same thing: They perform stronger among higher education groups, and do the best among those with postgraduate degrees. On the other hand, note that Trump consistently has the lowest level of education to himself.
Now, to reiterate, this doesn’t mean that Rubio and Kasich are splitting these voters. Rubio is actually performing quite a bit better among them, generally speaking, as reflected in his overall vote shares. It just means that when Kasich pulls in votes, they are coming from groups that are generally favorably inclined toward Rubio, and vice versa.
What about if we look in terms of income?
Again, this is not perfect, but we see something of the same tendency: Higher levels of income correlate with better showings for Rubio and Kasich, lower levels tend to go to Trump, while Carson and Cruz are somewhat inconsistent.
What about ideology?
Once again, this is not perfect, but note that Kasich and Rubio tend to be strongest among the same groups of voters.
Interestingly, Trump often runs strongest among moderates. This re-emphasizes a point that is crucial to understanding Trumpism – he is not what we think of as a “purification” Tea Party candidate. If you look at voters who feel betrayed by the GOP, Cruz tends to run strongly among them, rather than Trump. Trump is a more or less a pure populist, and that is a distinction with a difference.
Even on specific policy questions, we see something of the same split:
On the question of whether or not we should ban Muslim entry into the country (which was not asked in every state), Kasich and Rubio’s support is overwhelmingly concentrated among opponents of the ban, while the other candidates compete for voters that support the ban. A similar battery of questions were asked regarding support for deporting illegal immigrants versus offering a path to citizenship (perhaps surprisingly, the latter option won in every state), and similar splits occurred.
We’ve talked a lot about Rubio and Kasich, but as David and I noted Tuesday, Cruz and Carson tend to compete for the same sorts of voter. To see this most plainly, let’s look at whether voters consider themselves to be evangelicals, and how they voted:
Here we see that Trump’s support is more or less evenly spread, while Cruz’s and Carson’s coalitions are heavily drawn from evangelical subpopulations, and Kasich and Rubio tend to compete for non-evangelicals.
Now none of this is to say that Rubio entirely draws from Kasich, or that all of Carson’s support comes from Cruz voters. It is to say that these candidates draw disproportionately from one another, and that this matters in a close race. For example, without Kasich’s 4.2 percent support in Texas, Rubio probably clears the necessary vote threshold to win delegates there (and vice versa). Rubio lost Virginia by three percentage points; if he had won, say, two-thirds of Kasich’s 9.4 percent, he probably wins the state. Rubio probably prevented Kasich from winning Vermont (and vice versa).
On the other hand, Carson’s 6 percent may well have kept Cruz from closing the 2 percent gap between himself and Trump in Arkansas, and kept him out of second place in Georgia.
The point here isn’t that this candidate or that candidate ought to drop out, although Carson and to a lesser extent Kasich probably do not have a viable path to the nomination. What it does mean is that, to understand the dynamics of this race, you have to understand the dynamics of party factions. Because that is the engine that drives this train, for now.