Let’s Talk about Race and Bayes: How a British Statistician Helps Explain Racial Dialogue in America

Screen Shot 2016-03-10 at 10.29.13 AM

It often takes a little bit of distance from events before you start to get a feel for what they mean. The O.J. Simpson trial was 20 years ago. I was just old enough to remember it being an outsized cultural event – Judge Ito, along with Paul Harvey and Mo Vaughn, is on the list of vaguely familiar figures from my childhood. Twenty years, it turns out, is about the amount of time we needed to get an interesting fictional treatment of the three-ring circus that surrounded the murders of Nicole Brown Simpson and Ron Goldman, as well as the eventual acquittal of “Juice”.

Watching the FX series The People vs. O.J. Simpson with the benefit of hindsight, it’s hard not to be struck by the radically different views held by white and black Americans about the trial. Specifically, it’s tempting to wonder how in the world ‘the black community’ got it so wrong: at the time of the trial, about three-quarters of black Americans thought O.J. was innocent. (Today, there’s no room for serious doubt: he did it.)

The guilt or innocence of Simpson was obviously the marquee question of the trial, often overshadowing – at least in the public memory – the very serious missteps of the LAPD. As one contemporary observer put it, the simplest explanation was probably that the police had managed to frame a guilty man.

In any event, the questions raised by the trial, which the FX series captures nicely, are as relevant today as they were then. When Trayvon Martin, a black youth in Florida, was killed in 2012, there was a definite racial component to whether or not people thought the shooting was justified. The same thing happened when Michael Brown, also a black teenager, was shot in Ferguson, MO by a white police officer: the public’s opinion split pretty dramatically along racial lines.

When confronted with this pattern, most people hunker down and selectively appeal to whatever evidence they think bolsters their side. Why doesn’t everyone agree with your position? Probably, the typical person seems to think, because they’re too stupid or too flawed to see the truth.

Enter Thomas Bayes. An 18th-century English Presbyterian minister, Bayes is best known for formulating the theorem that bears his name. It’s an elegant idea with wide application and common sense appeal. It’s also, I think, crucial for understanding what we talk about when we talk about race.

Bayes Theorem: It's really quite simple.

Bayes Theorem: It’s really quite simple.

At heart, Bayesian analysis deals with how we interpret evidence. We can simplify and say that there are two competing schools of thought. The first school says that you try to gather the evidence in front of you, analyze how compelling it is, and then make the best decision you can about what you think is likely the truth. This is what most people think they’re doing and largely how public debates are discussed: the media pretends to present the facts, and the public pretends to decide on the basis of those facts.

The alternative model is Bayesian analysis, which comes down to this: you always have to interpret new evidence in light of prior beliefs. If you see a large bipedal animal covered in hair with abnormally large feet walking down the street, you might not be able to tell whether or not it’s Sasquatch just from what you’re seeing. But coupled with your prior beliefs and knowledge about Big Foot (namely, that no one has ever proven that they’ve seen him), you can probably conclude that instead it’s a guy in a costume.

I’m simplifying wildly, and there are really interesting debates about when Bayesian analysis is the preferred statistical method for hypothesis testing. (The medical field, for example, is a good candidate for Bayesian techniques.) But my focus here is less on championing Bayes’ theorem as better or worse than the alternatives, but rather in arguing that it’s more descriptive of how people actually reason about the world.

Most people adhere to Carl Sagan’s famous dictum: “Extraordinary claims require extraordinary evidence.” But the contrapositive is also true: obvious claims require little evidence. The rub, of course, in deciding what’s an extraordinary claim and what’s an obvious one.

To many Americans who watched the O.J. Simpson trial, or the George Zimmerman trial for the shooting of Trayvon Martin, or followed the the stories surrounding the shooting of Michael Brown, the claim that the police acted inappropriately is not an extraordinary one. The idea that Simpson, Brown, and Martin might have been singled out because of their race is, on the other hand, a pretty obvious claim. To other Americans, the opposite is true.

In 2015, Sandra Bland died while in the custody of Texas law enforcement after being pulled over for a minor traffic violation. The authorities contend that she committed suicide. Detractors from the official account imply, or outright claim, that she was killed by police. As is often the case with suicides, and as you would also expect if her murder were being covered up, the details of Ms. Bland’s death leave a lot of unanswered questions. How do people make sense of the explanation for why Ms. Bland was arrested, or the video that emerged of her interactions with a police officer, or her digital presence in the time leading up to her death, which seemed to show a woman filled with vigor?

A lot of it comes down to this: how extraordinary is the notion that police would kill an innocent person and then cover up the murder? In Bayesian terms, it comes down to your prior beliefs.

And that’s the major dysfunction in how we talk about race in America. Too often the conversation depends on the sort of episodic stimulus – a high profile crime or controversy – that encourages us to debate the facts but not actually drill down to our priors that influence how those facts actually fit together into a narrative. Whether it’s race, the police, public transportation, public schools, infrastructure, or taxes, Americans form their worldviews by appealing to their prior beliefs – beliefs shaped both by their unique experiences and by the experiences that characterize people like them. Being largely rational creatures, our priors are a reflection of our reality.

But we don’t all live in the same America – any conversation about our social challenges that doesn’t begin with that premise is going to go nowhere fast. If you want to get to the heart of disagreement, the first step is always to listen.

Follow Pedro on Twitter @IamPedroA.

Want to help The Fog of Policy grow? Then take a minute and share this piece! Or let me know what you think in the comments section.

Have a question or suggestion for a new piece? Submit it through the Feedback form – and don’t forget to subscribe on the homepage to get posts and features automatically sent to your inbox.

Leave a Reply