Donald Trump greets supporters in Mobile, Ala., on August 2015, during his campaign for president. (Mark Wallheiser/Getty Images)

To appropriate a phrase from the late Rick James, reflexive partisanship is a helluva drug. And today's Republican Party is much more united on what it is against — namely, the Democrats and the mainstream media — than on what it's for.

But when you ask a certain kind of question, you can see pretty clearly that GOP base voters aren't totally enamored of President Trump. They may stand by him, yes, but not necessarily because of him.

The annual Pew Research Center political typology survey asks enough different kinds of questions and slices it enough ways that you begin to see this. The study breaks down both right-leaning voters and left-leaning voters into four groups each, lumping them together using similar traits and views.

Here's how they break down along the political continuum:


Anchoring each end of the spectrum is the base — the most devoted and ideologically extreme partisans who almost always toe the party line on the issues of the day. These groups dominate primary electorates and have outsize influence because they are much more likely to vote and participate in politics. They are less than 30 percent of the country, but they are 45 percent of the electorate.

On the GOP side, this is what Pew calls the “core conservatives.” And this is the group in which you see some cracks in Trump's hold on the base.

Pew asked people whether they agree with Trump on “all or nearly all,” “many but not all,” “a few” or “no or almost no” issues. Even among the most conservative group, less than a majority — 44 percent — agreed with Trump on at least “nearly all” issues.

By contrast, 91 percent of the most liberal voters — the “solid liberals” — say they disagreed with Trump on nearly all issues.


Similarly, just 41 percent of the “core conservatives” group says it liked the way Trump conducts himself as president, compared with a majority — 51 percent — that said it had “mixed feelings” about Trump's conduct. And 8 percent said they outright didn't like it.


In both cases, you'll notice that more middle-of-the-road Republican-leaning voters were actually about evenly split on Trump's policies and conduct. They were divided between agreeing with him on at least “many” issues and just “a few” or less. More actually said they disliked his conduct than liked it.

But while these are less-reliable GOP voters, they do approve of Trump's job performance by wide margins, as the chart below shows. In other words: They do so despite not loving him personally or on policy.


Perhaps the most telling statistic in the bunch, though, is on image ratings -- and specifically how Trump compares with another Republican, Vice President Pence. Among the “core conservatives” group, 36 percent have a “very favorable” view of Trump personally, but significantly more — 50 percent — have a “very favorable” view of Pence.

Even among the next most right-leaning group — “country-first conservatives” — Pence leads Trump on “very favorable” views 45 percent to 41 percent.


And now look at the second part of that chart. It summarizes the real reason that these voters stand by Trump even though he's not close to their presidential ideal: their views of the alternative. Fully 8 in 10 of the most conservative Republican-leaning groups say they have a “very unfavorable” view of Hillary Clinton.

In fact, in nearly every group, twice as many in these GOP-leaning groups strongly dislike Clinton as strongly like Trump — 81-36 among “core conservatives,” 80-41 among “country-first conservatives,” 65-24 among “market skeptic Republicans” and 42-19 among “new era enterprisers.”


This is the great uniter. These voters know what they almost all agree upon. Trump may not be great on their policies, and they may even think he's kind of a jerk, but he's with them on the most important thing: being not-the-other-side.

It's arguably his most pronounced quality. And in an increasingly polarized country, it's what really matters.

Kevin Uhrmacher contributed to this post.