President Trump holds a rally at Olentangy Orange High School in Lewis Center, Ohio, on  Aug. 4. (Leah Millis/Reuters)

There are those things about the 2016 election that any American who’s managed to keep one eye open or who has opened a news site even once, even by accident, can recite by heart. President Trump lost the popular vote narrowly but won the electoral college. His base sticks with him and he’s very popular with Republicans, but his base is only about a third of the country, and overall, most Americans don’t like the job he’s doing as president. His base is white men without college degrees; the most fervent opposition he sees comes from black women and white women with college degrees.

People love to fight over these things: whether it matters that his base sticks with him, or the media’s role in affirming that base, or how the majority of Americans don’t like him, or how all of this affects the 2020 election — an election which, like the 2016 contest, exists outside of time itself. The 2016 election was almost two years ago, but it feels as recent as yesterday; the 2020 election seems as though it’s hopelessly distant and will be in front of us tomorrow.

Much of this, though, is anecdotal. We have exit polls and such that tell us what actually happened in the campaign, but many of the stories about Trump supporters standing by him ask individual people how they feel. “I stand by him,” those people say, speaking from behind the breakfast platter at Sal’s Diner in Newton Falls or while pointing out the broken window behind which they worked at the shuttered manufacturing plant in Mount Clemens.

Data released  Thursday by the Pew Research Center, though, offers something different: a look not only at who stands with Trump but who doesn’t — using data that is concretely tied to actual Trump voters.

The pollsters at Pew asked particular people how they felt about Trump in April, September and November 2016, as well as in March 2018. Those who said they voted for him and who were confirmed to have actually voted were categorized into four groups. Those who felt positively about Trump in both April 2016 and March 2018 were called “enthusiasts”: They liked him then, they like him now. Those who didn’t feel warmly about him in April 2016 but did in March are “converts,” since they came around to Trump eventually. And so on.

The breakdown looks like this.

The two groups at right, making up 18 percent of Trump voters, are particularly interesting. The “disillusioned” are people who found Trump appealing in April 2016 but who no longer do. The “skeptics” — 1 in 8 Trump voters — were ambivalent about Trump at both points, often voting for him, one assumes, because of a lesser-of-two-evils mentality. (Among the fifth of the electorate who viewed both Trump and Hillary Clinton unfavorably on Election Day, exit polling shows that Trump beat Clinton 47 to 30 percent.)

Where things get interesting is when we overlap demographic splits on those groups. An excellent New York Times report on Pew’s data noted that only a third of Trump voters were white men without college degrees and that those most likely to have lost faith in him were largely women and the college-educated. We already knew this, to some extent, given the shift among white, college-educated women in generic-ballot polling.

But that pattern is clear in Pew’s data.

More than half of both the “enthusiasts” and “converts” groups are men, that data shows. But 6 in 10 “skeptics” and “disillusioned” are women. (Because of the small sample size of the “disillusioned” group, it’s combined with “skeptics” in the data below.)

In addition to showing who fell away from Trump, this also reinforces where his strength lies: largely with men. Women were twice as likely to be “disillusioned” as men.

The pattern on education is similar. About three-quarters of his “enthusiast” base doesn’t have a college degree. About a third of each other group does. In other words, as we’d expect, it’s those without degrees who have been the most consistent base of support for Trump.

About 6 in 10 Trump voters who don’t have college degrees fell into the “enthusiasts” category; more than 8 in 10 are “enthusiasts” or “converts.” Only half of those with college degrees are “enthusiasts.” Of all of the demographic categories Pew provided to us — race, gender, education and generation — those with college degrees were the demographic with the smallest percentage that fell into the “enthusiast” group. Notice, though, that most of those in the disillusioned category are still people who don’t have degrees.

More interesting still are the data by race. About 12 percent of Trump “enthusiasts” are nonwhite. Among “converts,” though, those who gradually warmed up to Trump, only 6 percent are nonwhite. Among “skeptics” and the “disillusioned,” larger percentages are nonwhite.

None of this is exactly unexpected, but these specific data, tied to specific voters, give a sense of movement in Trump’s pool of voters that isn’t immediately obvious from other analyses. While most Trump voters stand by him, 12 percent never warmed up to him, and 6 percent lost the enthusiasm they once had.

The broad assumptions outlined in the first paragraph of this article are not inaccurate. But our tendency to accept those summaries as all-encompassing is worth exploring more often.