I would have thought we could have moved on by now. Both before and after the 2016 election, there were months of acrimonious debate over the question of whether Trump voters were motivated by racial hatred or anxiety over their economic prospects. And I thought the general conclusion would have been that the premise was wrong to begin with, that you couldn’t talk about “Trump voters” as a single unit, because the category includes a broad spectrum of people with a varying set of motivations. Some of them liked Trump’s rhetoric on jobs and globalization, some liked his rhetoric on immigration and Islam, and some liked all of it. Both of the appeals obviously contributed to his victory. (Those of us on the left, however, frequently suggested that Democrats should focus on winning over the economically-motivated Trump voters, rather than the wealthy racists, because the ones anxious about jobs are the ones whose support Democrats have a greater chance of peeling off.)

The “racism or economics” debate is a pretty easy one to resolve, then. Trump’s campaign was based on bigotry, but also fueled by a backlash to the unfairness of the contemporary globalized economy. And many workers fell for his promises to bring jobs back, just as racists got excited over his stigmatization of Mexican immigrants. A question that appears contentious and intractable actually has a fairly obvious answer.

But British journalist Mehdi Hasan has decided to reignite the debate once more, with a new column in The Intercept arguing that racism was the primary cause of Trump’s victory and that Democrats who say Trump voters were hurting economically are “trafficking in alternative facts.” Hasan is blunt and his conclusions unqualified: “The race was about race,” he says. “It’s not the economy. It’s the racism, stupid.” Hasan singles out Bernie Sanders and Elizabeth Warren for criticism, saying that by claiming Trump voters were economically motivated, Sanders and Warren are ignoring the “stubborn facts” and “coddling…those who happily embraced an openly xenophobic candidate.”

Hasan’s column repeats arguments that have been made over and over for two years, from Salon to Vox to The Atlantic. Many liberal pundits have consistently dismissed the idea that Trump voters acted out of defensible economic motives, instead suggesting that they were just as deplorable as Hillary Clinton made them out to be. (In fact, they go beyond Clinton, who was trying to draw a distinction between those who were deplorable and those who should be respected and listened to.) The position is somewhat surprising coming from Hasan, though, who has often seemed sympathetic to the Sanders left, and it’s doubly surprising for appearing in Glenn Greenwald’s Intercept, which has been consistently critical of Vox-ian liberalism.

If Hasan thinks this is true, then, it is worth dealing with his evidence. His argument for the proposition that the election was “about race” is as follows: There are a series of statistical correlations between racism and Trump support. Donald Trump did better than Romney or McCain among voters with high racial resentment. The best way to predict whether any given person is a Trump supporter is to ask them whether they think Barack Obama is a Muslim. If they say yes, they’re almost certainly a Trump supporter. (“This is economic anxiety? Really?” comments Hasan incredulously.) Those who hold negative racial stereotypes about African Americans are far more likely to be Trump supporters. (“Sorry, but how can any of these prejudices be blamed on free trade or low wages?”) On the other hand, having a low income did not predict support for Trump, and Trump supporters actually tend to have higher incomes than Clinton supporters. And while there may be “economic anxiety” among Trump voters, it tends to be the product of racial resentment rather than its cause; in 2016, people who were racist tended to be economically anxious, while people who were economically anxious did not thereby become racist.

These are the entirety of the facts that Hasan presents to support his conclusion that the election was “about” race and that Bernie Sanders is factually wrong to say things like “millions of Americans registered a protest vote on Tuesday, expressing their fierce opposition to an economic and political system that puts wealthy and corporate interests over their own.”

I have long been critical of those in the political press who loudly insist on their superior allegiance to Fact and Truth. By contrast with Hasan, who quotes John Adams that facts are “stubborn things,” I tend to believe facts are fundamentally slippery things. Statements that are literally factually true can often be highly misleading, and sometimes you do actually need the addition (not substitution) of some “alternative facts” in order to understand what is really going on. For example: I can cite GDP growth as proof that Americans are doing well economically. But it’s not until I understand the distribution of the economic benefits across society that I will know how the majority of Americans are actually doing. Or I can cite the fact that lifespans are increasing as evidence that American healthcare is “making us live longer.” But it might be that richer people are living longer while poorer people are actually living less long, making the word “us” erroneous. If a fact is true, but is incomplete, then it might actually leave us more ignorant than we were before.

This is precisely the situation with Hasan’s statistics. They are carefully selected to support his argument, with the statistics that don’t support it simply ignored. He, like many others who have written “it’s about racism” pieces, depends heavily on evidence that racism “predicts” support for Trump while income doesn’t, meaning that racists are more likely to be Trump supporters while poor people aren’t more likely to be Trump supporters.

But if we think about this statistic for a moment, we can see why it’s a dubious way of proving that Trump support was “about” race. First, Hasan is confusing the statement “Most racists are Trump supporters” with the statement “Most Trump supporters are racists.” Of course most racists are Trump supporters; racists tend to be on the political right, because the political left defines itself heavily by its commitment to advancing the social position of racial minorities. It would be shocking if racism didn’t predict support for Trump, because it would mean that racists had decided to ignore David Duke’s endorsement of Trump and vote for a candidate who embraced the language of “intersectional” social justice feminism. Nor is it surprising that Trump did better with racists than his more centrist predecessors. The more racist your campaign rhetoric is, the more the racists like you.

The income statistic is similarly unsurprising. Of course Trump’s supporters tend to be higher income. Republicans are the party of low taxes on the rich, and Trump wants to lower taxes on the rich. Democrats are the party of social programs for the poor. So poor people were always going to disproportionately be for Clinton, and rich people were going to disproportionately be for Trump. Furthermore, since Democrats are disproportionately the party of racial minorities, and racial minorities tend to be less wealthy than white people (due in part to several hundred years of black enslavement), the racially diverse Democratic base will ensure that poverty doesn’t predict Trump support.

Note how neither of these facts address the actual question. If we want to understand the relative role of race and economics in creating votes for Donald Trump, it doesn’t really help us to know that racists tend to be Trump voters. Imagine we have 100 voters, 10 of whom are high-income racists and 90 of whom are low-income non-racists concerned about the economy. Well, we know our 10 rich racists will probably vote for Donald Trump. And we know that being a low-income non-racist doesn’t really predict support for Donald Trump, so let’s say those votes split equally, or even break slightly in favor of Clinton. We count the votes, and the result is: 54 Trump, 46 Clinton. Trump gets 10 rich racists, plus 44 poor non-racists. Clinton gets 46 poor non-racists.

We can see, then, what can be concealed by statistics showing that “wealthy racists tend to support Trump” and “poor and economically anxious people tend to support Clinton.” Those two statistics are consistent with a situation in which the vast majority of Trump’s support occurs for economic reasons rather than racial ones. Yes, it’s true, the presence of racists in Trump’s coalition put Trump “over the top.” But it’s also true to say that the Democrats losing half of all economically anxious people put Trump over the top, and if you focused on the racism, you’d be focusing on the minor part of Trump’s overall support.

In laying out this hypothetical, I am not attempting to show that this is actually what happened. The two statistics (“racists support Trump” and “poor people support Clinton”) are also consistent with a situation in which 100% of Trump’s supporters are racist. Instead, I am demonstrating that the two premises in and of themselves can’t lead us to the conclusion Hasan wants to draw (and that other pundits have drawn over and over from them), which is that Trump’s support was about racism.

Hasan calls the idea that Trump “appealed to the economic anxieties of Americans” a fiction and concludes that “instead, attitudes about race, religion, and immigration trump (pun intended) economics.” But what he’s proved is that racial attitudes trump economics as predictors of a particular individual person’s support for Donald Trump, not that racial attitudes trump economics as the main issue Trump voters cared about or the main reason for his success. If we take the question “Was the election about race or about economics?” to mean “What was the relative role of race issues and economic issues in determining the outcome of the election?” then Hasan’s evidence does not actually address his question.

To get closer to a real answer, we might do better to look at what the most important issues were to Trump voters. What attracted them to Trump? Do they care more about economics or about race? We can begin to get an answer from a Pew poll conducted in July of 2016, which ranked issues by their importance to voters, broken down by the candidate they were supporting. Among voters generally, the economy was considered a “very important” issue to 84%, with immigration only the sixth-most important issue. Among Trump supporters, though, economic issues were considered very important to 90%, compared to 80% of Clinton supporters. For Trump supporters, immigration was the third-most important issue, with 79% considering it very important. Thus nearly every Trump supporter was “very” concerned about economic issues, and economic issues won out by at least 10% over immigration.

We still don’t know very much from this. But we do know that a good chunk of Trump supporters cared about economics without caring as much about immigration (and we must assume that all Trump voters who cared about immigration were racists in order to accept Hasan’s conclusion). Of course, “being worried about the economy” can mean a lot of things; a rich man can be worried about his tax rate increasing, and we don’t know anything about racial attitudes from this survey. But it should caution us against coming to simple conclusion like “the election was about race.”

Even if we stick to demonstrations of the factors that predict Trump support, we find Hasan burying crucial evidence. Hasan quotes a Gallup report that, in his words, “found that Trump supporters, far from being the ‘left behind’ or the losers of globalization, ‘earn relatively high household incomes and are no less likely to be unemployed or exposed to competition through trade or immigration.’” But let’s look at the original context of that quote:

[Trump’s] supporters are less educated and more likely to work in blue collar occupations, but they earn relatively high household incomes and are no less likely to be unemployed or exposed to competition through trade or immigration. On the other hand, living in racially isolated communities with worse health outcomes, lower social mobility, less social capital, greater reliance on social security income and less reliance on capital income, predicts higher levels of Trump support.

Hasan’s presentation of the Gallup analysis therefore borders on intellectual dishonesty. If you quote the bit about high average incomes and no lower likelihood of unemployment (facts which, as I explained before, we would expect given the general composition of the Republican base compared to the Democratic one), but you don’t quote the part about bad health outcomes, blue collar jobs, and low social mobility, then you’re selecting only those facts that confirm your worldview and refusing to deal with the ones that contradict it.

This is the trouble with Hasan’s overall argument, and with these types of pieces generally. They accuse others of ignoring “the facts,” but they don’t really care about facts themselves. Otherwise, why wouldn’t Hasan mention the fact that the economy was “very important” to 90% of Trump supporters? Why wouldn’t he even deal with that statistic, even if he had a good argument for why it should be disregarded? It’s the duty of a responsible political analyst to address the evidence that undermines their position.

Hasan is likewise unfair in his characterization of the Sanders/Warren position on Trump voters. He says that “for Sanders, Warren and others on the left, the economy is what matters most and class is everything.” But Sanders repeatedly accused Trump of running a “campaign of bigotry” and whipping up nativist sentiments. In the op-ed Hasan quotes, Sanders says that “millions” of Trump voters voted out of economic concerns. But he does not deny that large numbers of Trump’s voters may be racist. (He has explicitly acknowledged that “some are.”)

In fact, I don’t know a single leftist who denies that Trump ran a racist campaign that energized racist voters. The leftist position is, rather, that there are many (“millions of”) Trump voters who were drawn to his anti-Establishment stance because of their economic hardships, that Democrats should have had a better message to target those particular Trump voters, and that suggesting Trump voters as a unit are racist is both politically unwise and unsupported by evidence. Hasan is extremely derisive toward this position, with his repeated suggestion that it’s factually ignorant, even stupid. But he doesn’t offer any actual proof for why it’s wrong. Instead, he willfully mischaracterizes it.

Actually, the left-wing stance here should be extremely uncontroversial. It doesn’t even have to presume that the majority, or even a very large percentage, of Trump voters were “economically anxious” rather than racist. Consider the 100-voter scenario from earlier. Say we have 48 rich racists and 52 poor anxious people. Trump snags all the racists by default, but then manages to lure 4 anxious poor people through his message on trade. Trump wins. In that situation, it’s still worth pointing out that Democrats needed a better economic message, and that economics were an important determinant of the outcome. A lot of the misguided attempts to decide what the election was “about” result from failures to think about marginal differences. If most Trump voters were racist, and a minority were economically anxious, and the election was decided by a small number of votes in Rust Belt states (which it was), then politically you might reasonably decide that it’s not worth focusing on the racists (who will never vote for you) and instead you should craft a rhetorical appeal to the economically anxious Rust Belt voters who can mean the difference between winning and losing. (As I said, though, so much depends on how you want to define the phrase “what the election was about.” If it’s about majorities, you might get one answer. If it’s about margins, you might get another. In Trump: Anatomy of a Monstrosity I go into more detail about how anyone can construct any story they like about the election and have it be true in a certain sense.)

I should add here that the necessity of fairness applies no matter which side of this you think is correct. If I say “90% of Trump voters thought the economy was the most important issue, therefore the race was about economics,” and I do not mention or deal with the disproportionate amount of racial prejudice among Trump voters, I am also cherry-picking the facts that support my preferred conclusion. Anyone who tells you the one issue that the election was “about,” and cites factors that “predict” support, without telling you the full range of relevant information, is arguing either ignorantly or dishonestly. They are not putting all of the facts on the table; rather, they are just giving the evidence that supports their own position. This is partisanship and bias, which nobody should engage in. Having a well-defined set of political commitments does not justify misrepresentations of the truth.

Frankly, Hasan’s column saddens me. I have really respected some of the excellent work he has done on his interview programs (even though he has a consistently irritating tendency to constantly interrupt his guests). And I’m disappointed in The Intercept, which promised to follow Glenn Greenwald’s idea that you can be opinionated and honest at the same time, for publishing it. That’s not because it offers a conclusion I disagree with; I’m happy to have a discussion about the role of racism in the 2016 election, as weary as I am of that particular debate. Rather, it’s because Hasan uses the characteristic argumentation technique of the glib pundit: instead of helping the reader think through an issue and showing your work, you just throw out a few random statistics that back up your position.

The truth about race and economics in the election is easy to grasp. They both mattered, and we can focus on whichever we choose. (Personally, I think that means focusing whatever is most useful or instructive, and that the question “Do Trump supporters tend to be racist?” is less consequential than “Are there enough non-racist, economically anxious Trump voters to where economic anxiety played a significant role in his margin of victory thereby meaning Democrats need to address the issue more?”) And if Mehdi Hasan were as committed to Facts and Truth as he professes himself to be, he would be happy to concede this rather than perpetuating a pernicious misrepresentation.