George Zimmerman and the Power of Bias

So George Zimmerman apparently helped rescue a family from a car wreck. This sort of thing should only happen in picaresque novels, but apparently, in real life, George Zimmerman helped a family whose SUV had flipped after it went off the road. Right-wing blogs took this as confirmation that Zimmerman is just a good guy who was trying to keep people safe the night he followed, and eventually shot, Trayvon Martin.

And what about left-wing blogs? Did they update their opinion of Zimmerman's interior life accordingly? Not exactly, as far as I can tell. Folks like Max Read of Gawker hastened to emphasize that he was still an awful person. Over at Jezebel, editors speculated that he'd "camped out under an army blanket just waiting for a car accident at a busy intersection so he could begin to repair his damaged reputation."

In short, whatever you thought about George Zimmerman before, you still think of him now. That seems to be about par for this course; people formed opinions very early, and by a few days after the story broke, those opinions were pretty much inalterable. People claimed to be basing their opinions on the facts, but many of those facts were bogus. Either they were erroneous reports that had been disproved (Trayvon Martin beat up a homeless guy; George Zimmerman had no injuries and told the dispatcher that Martin was suspicious because he was black), or they were fairly extreme extrapolation of stuff that did happen.

For example, the dispatcher told Zimmerman that "we don't need you to do that" after Zimmerman said he was following Martin; after that, it's not clear whether Zimmerman continued following or not. In many narratives, it remains a known fact that Zimmerman continued following Martin in direct violation of police orders, even though the dispatcher testified that this wasn't an order, and they have no authority to give out such orders.

On the other side, the fact that Trayvon Martin was still walking around outside 40 minutes after he could have been warm and dry at home has been transformed into the certainty that he was up to no good, rather than, say, wandering around while talking on the phone, something that I do pretty much anytime I make a call. People talk about Martin sucker-punching Zimmerman, or Zimmerman stalking and attacking Martin, as if these were established facts rather than highly contested speculation.

But in a way, even more remarkable are the stories that people tell about how the confrontation went down. I heard a lot of commentary to the effect that Zimmerman was guilty, morally if not legally, because he was a "wannabe cop" and "vigilante" who got out of that car because the gun on his hip made the whiny little coward feel all powerful and manly, and he knew that he could always take refuge in Florida's Stand Your Ground Law if he shot some poor black kid. From conservatives, I heard that Martin was a "wannabe thug" who decided to take down the "creepy-ass cracker" for disrespecting him. These stories were inevitably presented not as possibilities, but as facts, as serious as cancer and as real as next week.

This is absurd. It's not that these stories aren't possible; they certainly are. I can easily imagine someone feeling safer, more willing to risk a confrontation, because he had a gun at his back. And I can easily imagine a 17-year-old boy deciding to haul off and hit someone who he felt was disrespecting him; I've seen it happen with my own eyes.

But my ability to imagine such stories doesn't mean that they're true. It is not even true that the thing I find easiest to imagine is the thing that is the most likely to have happened. What I find easy to imagine tells you a lot about me. It does not tell you all that much about other people, or the world.

Parents find it easy to imagine their child being kidnapped by a stranger, which is why many children under the age of 12 or 13 are now escorted everywhere by a parent or another trusted adult. But stranger abductions are incredibly rare and always have been, even in the days when first-graders regularly walked themselves to school. Parents find it easy to imagine their children dying in a gun accident, which is why you hear about parents who won't have guns in the house, and refuse to let their kids play at the homes of parents who do. But those sorts of accidental shootings involving young children are about as rare as stranger abductions. On the other hand, very few parents would say "I won't let you play at their house -- they have a swimming pool," even though drowning is one of the most common ways for young children to die. Economist Steven Levitt estimates that swimming pools are about 100 times more dangerous than a gun in the home.

Why do we overestimate rare risks, and underestimate common ones? Because the rare ones make the news. When a child is abducted, or shot, it is certain to make the evening newscast. On the other hand, "child drowns at home after falling into pool (or filled bathtub, or bucket)" is a personal tragedy, not a news story. We are led astray by "availability bias": We think that things we can call to mind most easily are also the most likely.

This is particularly problematic when we are imagining peoples' states of mind, because we're just not very good at it. We're prone to "egotistical bias," where we assume that someone who disagrees with us must be actively rejecting the values we hold dear, rather than pursuing some alternative value -- so pro-lifers say that pro-choicers don't care about babies, and pro-choicers retort that pro-lifers hate women's autonomy, and women. We make fundamental attribution errors, overestimating the role of personality, and underestimating the role of the situation -- so if someone cuts you off in traffic, you don't think "maybe his wife left him this morning and he's really distracted," you think "what a jerk!"

We are prone to "hindsight bias" -- we find it difficult to shed our knowledge of what has happened, so we tend to judge people based on the outcome, rather than what they could have known at the time. Both Presidents George W. Bush and Bill Clinton were excoriated by partisans for getting intelligence presaging the Sept. 11 attacks, and failing to act -- even though this intelligence was a few drops in a large bucket of ambiguous information about possible events that mostly didn't happen.

While I was writing my book, I went to visit the experimental economics lab at Chapman University, where I spent several days just talking to experimental economists. The most valuable thing they taught me?

"People are heterogenous," as one researcher told me. I heard it over and over. People are surprisingly rational in good market structures, but they're also frequently unpredictable. "Every experiment I have done has surprised me somehow," said Bart Wilson, who has spent two decades running economics experiments.

Consider this graph, which illustrates how individuals in experiments made decisions under uncertain conditions. In an economics textbook, we speak of "utility functions" and so forth, but Nathaniel Wilcox's experiments showed 80 different functions for making decisions, all slightly different.

Source: Nathaniel Wilcox

Imagine yourself as an economist, trying to predict what one of these individuals would do. You can't; each of them has their own individual process for making decisions. It's internally consistent. But there are clearly at least four different internally consistent ways to approach these problems, and then a lot of individual variation in exactly how those approaches are applied.

Now imagine one of the optimists looking at the behavior of one of the prospect theorists. You don't know that there are multiple rules. At first his behavior looks normal. But then he starts acting weirdly. How can we explain this? Did something happen? Does he have some nefarious purpose?

No, he's just different from you. But it's very hard to imagine someone else's thought process when it's not like yours; in fact, it's hard even to imagine that it's not like yours.

Often when you see decision science papers reported, you see people say things like "people are irrational; they ..." followed by "place more emphasis on the present than the future" or whatever the cognitive bias paper of the week says. But this reporting is misleading. "People" don't act in groups. Those papers are average, and they usually contain a great deal of variation, from people who barely discount the future against the present, to folks who can't wait 10 minutes to get their hands on a treat.

Here's my opinion of what happened on that dark night in Sanford, Florida: I don't know. I don't know what Zimmerman and Martin were thinking. I don't know who said what to whom. I don't know who threw the first punch, or why Zimmerman got out of his car, or why Martin didn't go home.

What I know is that we find it very easy to imagine the worst of people who are not like us. And we frequently confuse our imaginations with reality.

This column does not necessarily reflect the opinion of Bloomberg View's editorial board or Bloomberg LP, its owners and investors.