When New York Gov. Nelson Rockefeller ran for president in 1964 (as a liberal Republican, which was a thing then), the fact that he had been divorced and remarried (to a younger woman!) was treated as a potential deathblow to his presidential aspirations. Back then, more so than now, getting divorced was generally treated as a moral failing.

Rockefeller did end up losing the 1964 Republican presidential nomination to Barry Goldwater, and his divorce was often cited as a major factor. (The first divorced person to be elected president was Ronald Reagan in 1980.)

Then along came the thrice-married divorce-a-holic Donald Trump (each wife younger than the one before). And the (formerly) moralistic Republican Party forgave him his marital track record and even forgave him for bragging about grabbing women by the pussy and a great many other behaviors that were deemed disqualifying moral failings in earlier times.

Overwhelming evangelical support

Even more unprecedented given all of the above, Trump was elected with overwhelming support from evangelical Christians. I say it’s unprecedented based partly on polling data taken from the latest great column by Thomas Edsall, who basically explores the ironies and anomalies of it.

Before I go into the Edsall argument and findings, I just want to mention that the most unwaveringly loyal Trump supporter, since the day Trump put him on the ticket, has been Vice President Mike Pence, whose signature line has for years been (and still is), “I’m a Christian, a conservative and a Republican – in that order.”

I’m no expert on Christianity, but as best I understand it, Donald Trump is pretty weak on the Christlike virtues. He’s greedy, arrogant, lascivious and he definitely doesn’t turn the other cheek.  He’s also not much of a conservative, nor a Republican, as I understand those terms, except in the sense that he decided to seek the Republican nomination so he could be president. And he somehow obtained that party’s nomination.

Having done so, he obtained the support of the overwhelming majority of Republicans on Election Day, including many who are more serious about their Christianity and about their conservatism than the man for whom they voted.

It’s not about policy views

So Edsall’s point, if I could paraphrase it, is that if you think that people are Republicans (or Democrats) because they hold certain policy views, you are seriously misunderstanding. If you think people voted for Trump because of the policy views he held, with which they are agreed, or his personal moral qualities, of which they approved, you are also mistaken.

Edsall’s argument, which I hate to agree with but can’t help it, is that you will understand voter behavior better if you think of partisanship as an element of identity, and a powerful element. We are used to thinking of a set of elements of identity, like race and religion and class. We have come to accept that these elements of identity may be powerful predictors of voting behavior. Most blacks vote for Democrats. Most evangelical Christians vote for Republicans.

But if partisanship is also an element of identity, then it’s not too surprising that most, almost all, Republicans don’t just feel like Republicans as some ethereal aspect of identity; they also vote for Republicans. And the same goes for Democrats. In fact, most voters end up identifying with the party that their parents supported. I know it’s true of me and most people I know. There are surely many exceptions. But for most of us, rather than making a “choice,” based on either some of our political ideas or issue positions, it mostly works the other way around. Most of us grow up in a party, the party of our parents, and that informs our political positions on many issues, and we vote for that party most of our lives. (There are some who argue there’s a genetic component to this.)

These are not iron laws, but the correlation is pretty strong.

Identity in internal conflict

Now (and I’m really getting close to the big Edsall aha! moment) what happens when elements of your powerful partisan identity come into conflict with other powerful elements of your identity? You’re an evangelical Christian, which makes you more likely to vote for Republicans, but you are also more likely to care about conventional moral behavior. Or at least you were.

Edsall built his column around an interesting poll finding by a group called the Public Religion Research Institute, which sponsored a poll in 2011 that asked whether a political candidate who is immoral in private life will still behave ethically and fulfill their duties in their public and professional life.

In 2011, white evangelical Protestants, by 61-30 percent, were the group most likely to say that someone whose personal behavior was immoral would lack the ability to behave ethically in public life. All religious groups said no, but by much smaller margins. (If you’re curious, Catholics also said no, but only by 49-42; mainline (non-evangelical) Protestants said no, but only by 44-38 percent, and those without a religious affiliation said yes, by 63-26.

Frankly, isn’t that about what you would expect, that certain kinds of voters, like evangelical Christians for example, would drive a harder bargain on “moral conduct” than some other kinds?

But then, you would ask yourself: Most of those evangelical Protestants who answered that way ended up voting for Trump, whose personal ethical/moral conduct stinks. Why would they get behind someone whose immoral private life would render him unlikely to ethically fulfill his responsibilities as president?

Well, they didn’t, because in the year that Trump was the Republican nominee, their answer to that question moved – by 42 points!

The 42-point swing

Yes, that’s the big aha moment from the Edsall piece. A 42-point swing. 42! Evangelicals, who, pre-Trump, said that by 61-30 personal/private moral failings would be a strong predictor of public moral shortcoming in a candidate for president, in 2016, by 72-20 percent, now say that if a candidate’s personal behavior is immoral, that is no reason to assume that his or her public behavior will be also.

The other groups in the poll moved slightly in the same direction. But their movement was dwarfed by the movement among evangelicals, who went from being the  group that was most concerned about supporting a personally immoral candidate to the group least concerned.

It’s just one quirky poll question, but it’s hard not to be impressed. What to make of it?

Edsall pretty much offered an explanation, to which I referred above. His piece was headlined: “Trump says jump. His supporters ask: How high?” Which may err on the rude or crude side, but loops back to the point above, which I stole from Edsall and he stole from several political scientists quoted in the piece. It goes something like this:

The internal dialogue

If you accept that partisan affiliation is a piece of identity, it’s hard to go against one’s identity in deciding whom to support for president, but if different elements of one’s identity are in conflict, one’s mind or gut may look for a way. Imagine the internal dialogue of evangelical Protestant Republican needing to vote for Trump without willingly helping to elect an unethical president. Something like (I’m making this up):

“I used to believe that an immoral person cannot make an ethical president. And I can’t disregard the evidence that Trump’s life does not demonstrate moral behavior in his personal or business life. But I’m a Republican and cannot seriously consider voting against my party. So I have to adjust something, and the least painful thing is to adjust my belief that an immoral person cannot make an ethical president.”

Edsall asked the rhetorical question: “Are the moral convictions of white evangelical Protestants writ in stone?” And he answered: Apparently not.

If you read the full Edsall column, you’ll get a bunch of really smart political scientists weighing in on how and why this works.