Donald Trump is president because he won about 78,000 more votes than Hillary Clinton in Michigan, Pennsylvania and Wisconsin. He’s also president because he beat her more soundly in Florida and Texas and so on, sure, but it’s those 78,000 votes that have captured the public’s imagination. Why? For one thing, those states collectively had backed the Republican precisely zero times in presidential contests from 1992 to 2012. Moreover, those states reinforced the argument that Trump had made from Day One: Disaffected working class-white voters in the Rust Belt would put him over the top. Polling missed this surge and so, too, did media observers like yours truly — leading to a barrage of “

what’s going on with working-class whites?” analysis.

The Atlantic has entered the fray on the strength of a poll conducted in partnership with PRRI. Their analysis suggests that cultural anxiety, not economic stress, pushed those voters to endorse Trump. This agrees with exit polling, which showed that those most worried about the economy preferred Clinton in all of those states — and in most other states, too.

But this graph from PRRI’s analysis is worth breaking out separately. It suggests that younger working-class white voters — defined by PRRI as those without a college degree who don’t have salaried employment — are more strongly Republican than older members of that group.

Because party identity is generally fairly static over time, that graph suggests a potential long-term problem for a Democratic Party that received the strong support of younger voters for Barack Obama’s victories in 2008 and 2012 — and that seemed likely to be able to count on support from those voters over time. As I wrote in March, millennials — defined broadly and inconsistently as those born from 1980 to 2000 — tend to lean left, while older voters tend to lean more heavily conservative. This PRRI poll seems to add an asterisk, perhaps along the lines of this 2014 analysis: the economic struggles felt by these young people under a Democratic president might turn them off over the long term.

But Republicans shouldn’t celebrate just yet. First of all, this is a very small group of voters. And, second, PRRI’s numbers don’t match other polls.

Let’s address the second part first. PRRI’s poll used a very particular definition of white working class that’s a bit tricky to replicate. Often, pollsters use college degrees as a marker: Those without a four-year degree are determined to be in the working class, given how those degrees correlate to income.

When The Post and ABC News polled shortly before the election, whites younger than 40 and without a college degree were more likely to say they were Republican than Democrat — but by a much smaller margin than other age groups. (They were also far more likely to offer some other response besides Democrat, Republican or independent.)

The General Social Survey, a biannual national survey funded primarily by the National Science Foundation, gives us a bit more detail. Although the 2016 survey doesn’t break down income by type (hourly vs. salary) as was done in years past, we can get a more fine-tuned look at the group under consideration.

The survey offers a seven-part scale for partisanship: Strong Democrat, Democrat, independent who leans Democrat, independent, independent who leans Republican, Republican and strong Republican. (It’s important to remember that, broadly, there are more independents in the United States than Democrats or Republicans — but that most independents still tend to vote on partisan lines.) If we clump those results together into Democrat, independent or Republican, the result is as follows.

Younger whites without a college degree are much more likely to say they’re independent than older working-class whites — but are not any more likely to say that they are Republican.

The GSS also asks people to self-identify their economic class. Among young whites who identify as “working class,” the split isn’t much different from those with no degree — or from young white voters overall.

If, however, we include those independents who lean Democratic or Republican with those who identify specifically with a party, the results change a bit. Here, 19- to 29-year-old whites without college degrees are more likely to be strong Republican, Republican or independent-leaning-Republican than those ages 30 to 44. But most still identify with the Democrats, far more so than older whites without degrees.

That’s not the case with whites who self-identify as “working class.” In that group, those younger than 30 are more likely to identify as leaning Republican than Democratic (although still less so than older whites). Since this is a self-identification, there’s likely some overlap here with partisan identity. (For example, 15 percent of those making $100,000 or more view themselves as “working class” or “lower class.”)

Back to the broader point: The PRRI assessment seems to be something of an anomaly. That doesn’t mean it’s incorrect; it just means that we should be cautious about making detailed assumptions based on these figures.

Among the assumptions that we should be hesitant to make is that this bodes poorly for the Democrats long-term. Not only because the poll numbers aren’t replicated in other surveys, but because whites without college degrees are a smaller part of the young population than the older population.

The Census Bureau has tracked educational attainment for decades. Over time, the number of Americans with college degrees has increased sharply; more Americans today have degrees than at any point in the past.

What’s more, younger Americans are more likely to have a degree than older Americans — meaning that a smaller percentage of that age group likely fits into PRRI’s definition of “white working class.”

College degrees and economic success don’t necessarily go hand-in-hand, as anyone with outstanding college debt can tell you. Nor is it the case that having a college degree will continue to be a strong predictor of economic class. That so many more Americans will have college degrees in the future will shift the significance of this metric.

For now? The broad pattern still holds. Young people are more supportive of the Democrats than the Republicans. But the results of 2016 should be reminder enough for the left: That’s not necessarily going to be good enough.