Smart students, stupid charts
I stand behind everything I wrote in I Smell a Chart, at least until someone shares it with FiveThirtyEight’s Leah Libresco who will politely explain why I’m an idiot. My post focused on the abuses of data and statistics in the article, but I neglected an important point: if your data shows that affirmative action doesn’t help Hispanic students at all, but you still claim that it does, you’re doing actual harm to Hispanic students.
Affirmative action by definition evaluates people based on factors other than academic qualification, so employers who only care about qualification will try to counteract the effect of affirmative action. Let’s say that a university will accept a person named Leigh if their SAT is above 600, but they’ll only take Li if he scores 650, while Luis needs a mere 550 score. When an employer receives job application from all three students, they can conclude that Li was likelier to have a higher SAT score than Luis, which will influence their hiring decision. Even if they don’t aim to discriminate, companies in a competitive market have a huge incentive to hire the most qualified people they can.
I don’t know how large this effect is, but it’s not zero. It also depends on the perception by employers of affirmative action, not the actual effect of AA. If some people face severely limited educational opportunities and AA helps them get to college, AA may well have a positive effect even when rational employer discrimination is accounted for. But if AA doesn’t help Hispanics get into college, and the perception of AA hurts them coming out of college, it causes them net harm. Articles that increase the perception of AA while doing nothing to actually make it more effective exacerbate the harm even further.
Trump in the cockpit, journalists in the seats: story of a plane crash
With all that said, the AA article is an outlier on FiveThirtyEight. On the whole, the site is consistently smart, informative, and more or less objective.
To wit, FiveThirtyEight is practically the only mainstream (i.e. left-leaning) outlet that I can stomach following election coverage on (right-wing sources are no better). A lot of mainstream media companies chose to self-immolate in a Trumpnado of journalism malpractice. I was a loyal reader of Slate for years, until August 1st. On that day, they devoted 10 of the 12 articles on their homepage to attacking Trump, including “Trump Eats Fried Chicken Like a Sociopath”. I deleted the slate.com bookmark in disgust.
People have argued that it’s OK for journalists to follow their conscience and sacrifice their hard-earned credibility for a political goal they think is important. Fair enough, but guess what? Now these journalists are fresh out of credibility. The “fried chicken” guy writes almost exclusively about Trump, but who’s going to trust him on the topic now?
The worst part is, I really believe that Slate and NY Times and the rest of them are helping Trump, not hurting him. Contra the Donald, it says nowhere in the election rules that the media owes him fair coverage. If you’re unfit to be president, people will report that you’re unfit to be president. If you’ve assaulted women, the media will probably mention that. But when an outlet devotes a week to “Trump kicked crying babies out of a rally” (which didn’t actually happen), who’s going to take their word or even notice when they try to report serious news?
I have nothing but contempt for people who spit on truth-seeking the moment the truth becomes the slightest hindrance to the pursuit of their political goals. This goes both for presidential candidates and for journalists.
Speaking of Trump, do you remember when I encouraged everyone not to vote?
My recommendation was based on four premises:
- Mathematically, your vote is really unlikely to swing an election.
- People judge candidates on personality more than on policy, and personality shouldn’t matter.
- The actual policies enacted are unpredictable and will (on expectation) be very similar no matter who gets elected.
- Being engaged in politics will turn you against friends and family, the miniscule chance of a positive public impact not worth the personal cost.
Point #1 still stands. If you want to know your exact chance of tipping the election you can adjust the baseline 1-in-10-million chance using FiveThirtyEight’s Voter Power Index for each state.
On the other hand, points #2 and #3 were dispatched quite decisively in the 10 months since my post. Personality doesn’t matter when the choice of personalities comes from a pool of basically respectable people like Bush, Gore, Kerry, McCain, Obama and Romney. Trump is an extreme outlier, and it’s no coincidence that 0 of the 6 people mentioned in the previous sentence are on his side. On policy, I don’t know what’s scarier: that Trump, if elected, will actually enact his proposals on immigration, trade, taxes, NATO and abortion, or that he really has no ideology and will make shit up as he goes along. I’m with Sam Harris on this one: the important part of picking the lesser evil isn’t that it’s evil, what’s important is that it’s less evil.
But point #4 is the important one for each person to realize individually: that you pay the price for political engagement by letting hate into your heart and discord into your relationships.
I followed the 2012 election from a graduate program that was about evenly split between Obama and Romney voters. This caused many of my classmates to get angry at each other, but at least they sometimes talked. That’s why I recommended that people match up across party lines and go bowling on election day, they’d have the same impact on democracy but without losing half their friends.
In 2016, there’s no one to bowl with. It’s almost impossible to imagine a social or professional group that splits 50-50 between Trump and Hillary. The two sides have consolidated their bubbles, purged all thought enemies, and have nothing but disdain for each other. I’ll give a blue tribe example because that’s the tribe I live in, and a controversial example because it has to be controversial to make the point.
Four weeks ago, the blue tribe declared total war on Scott “Dilbert” Adams for predicting that Trump will win the election.
On 9/30, Slate (who else?) wrote a hit piece on Adams calling him a “horny narcissist” in the article’s title. The next week Adams was shadowbanned on Twitter, a company that seems to have abandoned its ideals at the first sign of financial hardship. In the weeks since, an avalanche of hate piled up on Adams from online mobs and the media.
By now, a month of unrelenting abuse has curved the trajectory of Adams’ posts towards the crazy end. But when I browsed his blog in early October, he seemed at worst like someone who came up with a half-true hedgehogian theory (the “Trump is a master persuader” hypothesis), and stuck with it too long through a combination of doubling down on a bet and plain old confirmation bias. If it’s a crime to have an ego and an online platform, let him without sin cast the first tweet. But Adams’ crime is triple: an ego, a platform, and an association to Donald Trump.
I don’t have any reason to sympathize with Adams besides empathy for a person targeted by a bullying mob. But this post isn’t for Adams, it’s for the 100,000th person who tweeted “you’re dumb and your comic sucks too” at him.
43% of Americans support Trump. That’s 137 million people. The vast majority of these hold more extreme political opinions than Adams did a month ago when he declared his main concerns to be Hillary’s health and her estate tax plan. If you declare Adams to be an unredeemable enemy, an evil alien whose psychology is incomprehensible to right-thinking folk, the same goes for at least 100,000,000 other Americans.
Unlike a political argument with your uncle or classmate in 2012, these 100 million in 2016 will never get the chance to convince you of their humanity. The tribes barely intersect as it, and now Trump supporters see that nothing comes from interacting liberals except derision and harassment. 100 million people now hate and fear the blue tribe. And what really sucks, most of them will also hate and fear people like me and Scott Aaronson, even though we write blogs in defense of Trump defenders’ right to defend him.
Scott and I are educated coastal Jewish liberals. Trump supporters don’t read our blogs, but they can tell we’re from the tribe of their enemies. And I really don’t want to live in a country with 100 million who see me that way. So that’s why I defend Scott Adams’ right to be an idiot online. Because those who attack him drown out the voices of reason and tolerance, the voices of actual liberalism. I’m a liberal, and I can’t stomach that.