Liberals want reality to conform to their heroic narratives about life. So wouldn't it be easier if Trump's support were mostly explained by racism (which is repulsive), rather than his supporters' status as low-earning, working-class people (which is sympathetic)? The reality is that these issues may be connected.

Because some liberals find explanations of Trump's support by "class" or "economic anxiety" to be prissy or evasive, they make a joke whenever a Trump supporter is found doing something heinously racist. They point to it and say, "Wow, lots of economic anxiety here." Naturally, their argument would be bolstered if it turned out that Trump supporters were discovered to not be so strained economically. Data journalist Nate Silver provided what looked like a reason to believe this. And, even though the evidence is quite thin, wonkish liberals are running with it.

Matt Yglesias, using numbers from Silver's article "The myth of Donald Trump's working class support," writes that Trump supporters are "pretty rich."

As best we can tell from the data available in exit polls, the median household income of a Trump supporter is about $72,000 a year. It's true that this makes Trump voters more downscale than John Kasich voters ($91,000 a year), but it's essentially equal to the median household income of Ted Cruz voters ($73,000 a year) and well above the $61,000-a-year median household income of Hillary Clinton and Bernie Sanders supporters.

Note that the median household income in the United States is only $52,000 a year and most people don't vote in primaries, so all of the major 2016 candidates turn out to have supporters who are more affluent than the typical American. Trump, in particular, built his big primary wins on the backs of people who are economically comfortable. There is no country on Earth where the median household income is higher than the median household income of a Donald Trump primary voter, and there never has been throughout world history. [Vox]

Yes, at this scale, you do notice that Trump voters are richer than pharaoh's slaves or 11th-century Russian serfs. I suppose they really ought to check their privilege. By this standard the poorest of America's poor are wealthier than many of history's kings and aristocrats. But let's just price in the fact that in the 21st century, dentistry and agriculture are as good or better than ever. Unfortunately, after you factor in the century and country we live in, the evidence that the Trump voter is especially well-off is pretty thin.

I'm a defender of data-journalism generally and Nate Silver specifically. But Silver should have been more explicit about the limits of his data set.

Household income numbers are not all that revealing. If you wanted to guess whether the residents of certain houses were obese, you'd want to know more than the total weight of human flesh domiciled in a given house. You'd want the divisors, like number of humans. You'd want other qualifying data, like their ages. The very numbers that would tell us most about the relative strain on a household income isn't available in the exit data Nate Silver uses.

We know that voters generally have higher incomes than non-voters, that primary electorates have higher incomes than general election voters, and that Republicans tend to have higher household incomes than Democrats. But data from previous elections also tells us that Republican voters tend to be older than Democrats. They are more likely to be married and more likely to have more children. And researchers like George Hawley have shown that rising income doesn't correlate with rising support for the Republican Party if income gains are accompanied by higher costs of land or the need for greater educational attainment.

Telling us the household income of Trump supporters without describing the age and number of the earners and dependents tells us not much at all. A 24-year-old single and unmarried Bernie Sanders supporter who makes $28,000 a year may have a college degree and may expect to make a six-figure income by middle age. He may have parents that help him make rent when he comes up short.

But under household-income numbers alone, that Sanders supporter looks more desperate than a Trump supporter who heads a family of five, in which two working parents in their peak earning years make $68,000 as a household. Sometimes being "rich" means you can spend a few years earning $24,000 at a job you love. Sometimes being working class means taking a dangerous job at six figures, but go on disability in your mid-40s. Pay gaps between America's classes are small early in the life-cycle of a career.

And so this raw data leads intelligent journalists to say insane things, as when Matthew Yglesias described the Trump-voting Staten Island as an "affluent" community.

When people think of "affluent" they think of neighborhoods they'd move into if they became wealthy. In New York that might be the Upper East Side or a bedroom community in Westchester County. Staten Island has a couple of wealthy neighborhoods on the North Shore. But the borough is defined by a culture of New York's public sector workers — like sanitation men, NYPD, and FDNY — whose incomes look "affluent" by national standards, but don't feel like it in such close proximity to Manhattan.

More truth is found out about Staten Island on Urban Dictionary, or in the character of Ray from HBO's Girls, who explains Staten Island this way: "All these people, they want to live in Manhattan but they end up on this f--ked up weird little island watching the city in the distance with this quiet–just–rage burning in their hearts." Yes, there are poorer neighborhoods in New York that vote Democrat. But Staten Island is not the Upper East Side.

The Staten Island example fits with some of the other data that Silver, Yglesias, and others breeze over. Staten Island was Trump-central, while truly affluent neighborhoods voted for John Kasich. Even though his campaign was a joke by this point, Kasich did best against Trump in truly affluent New York suburbs, like Darien and Westport, Connecticut. Trump dominated with voters in "affluent" Maryland that were over 50 or had no college degree. He did best with people reporting less than $100,000 in total family income. Nationally, Trump's household income numbers are matched roughly evenly with Ted Cruz's. But Trump was winning those household incomes in states with higher costs of living. Cruz dominated Texas, Wyoming, the inner Plains states. Trump won people with the same income in much more expensive states like Maryland and Massachusetts.

I would not argue that Trump supporters are the most put-upon subset of voters. They don't face the same poverty rates as other groups that vote for Democrats, nor the incarceration rates. And Trump does have many truly affluent, secular supporters in the Northeast. But the data Nate Silver provides does not prove the case that his working-class support is "a myth."

In the other ways we measure class besides income, a core of Donald Trump supporters sure do look like the picture of Fishtown drawn by Charles Murray in Coming Apart. Silver admits Trump supporters are notable for not having gone to college. Other surveys show Trump supporters, consistent with other downscale whites, do not regularly attend church. Jobs in agriculture, construction, and manufacturing correlate with Trump support. So does owning a mobile home. Trump dominates the "great white ghetto" that is Appalachia. We see the same in the admittedly anecdotal evidence that Trump supporters were practically "invisible" to New Hampshire's Republican establishment before they showed up at the polls.

My hunch is that even though Trump's support is now broader than the white working class, the white working-class portion of the GOP formed a crucially important core of his support throughout the primaries. And that these voters have a unique bond with him, one similar to the bond of business-class Republicans with Mitt Romney. We're going to have to wait for the micro-data to come out after the election to pronounce more definitely on the nature of Trump's support.

It is important to get this right. My own view is that explicitly racially-charged politics and nationalism are not merely relics of the past, but our likely future. And that racism can be generated or intensified when it is accompanied by real economic anxiety. If you perceive that your life prospects would improve in a tighter labor market for low-skilled workers, you may come to resent the policies that invite more low-skilled workers into the country. Motivated reasoning could also lead you to hate and fear the immigrants themselves for other reasons, however irrational.

Some residents of New Orleans moved from the reasonable belief that Republican government wasn't solicitous of their interests, to believing crazy things like George Bush and his friends blew up the levees to drown them intentionally. It would have been wrong to say, "They're just pathological, they can't be helped," and strike the city and its problems from the list of political issues it is permissible to address. The correct response is to create a policy environment that builds trust with the residents of that city again. I actually learned that ethic from liberals.

Similarly, if you want working-class whites to stop believing that immigrants are the cause of their economic problems, the best way for our government and political class to prove it is to create a nation in which a hard worker generally provides a better life for his family than his father did, and halt the trend of downward mobility among such a large portion of Americans. The political class should do this not because Trump voters meet their moral standards, but because they are fellow citizens. Or we could just condemn them as racists who are lucky to live in beautiful, "classy" Staten Island, and see how that works.