To give you just one sense of the mastery of Mullainathan’s machine, it flagged 1 percent of all the defendants as “high risk.” These are the people the computer thought should never be released prior to trial. According to the machine’s calculations, well over half of the people in that high-risk group would commit another crime if let out on bail. When the human judges looked at that same group of bad apples, though, they didn’t identify them as dangerous at all. They released 48.5 percent of them! “Many of the defendants flagged by the algorithm as high risk are treated by the judge as if they were low risk,” Team Mullainathan concluded in a particularly devastating passage. “Performing this exercise suggests that judges are not simply setting a high threshold for detention but are mis-ranking defendants.…The marginal defendants they select to detain are drawn from throughout the entire predicted risk distribution.” Translation: the bail decisions of judges are all over the place.
I think you’ll agree that this is baffling. When judges make their bail decisions, they have access to three sources of information. They have the defendant’s record—his age, previous offenses, what happened the last time he was granted bail, where he lives, where he works. They have the testimony of the district attorney and the defendant’s lawyer: whatever information is communicated in the courtroom. And they have the evidence of their own eyes. What is my feeling about this man before me?
Mullainathan’s computer, on the other hand, couldn’t see the defendant and it couldn’t hear anything that was said in the courtroom. All it had was the defendant’s age and rap sheet. It had a fraction of the information available to the judge—and it did a much better job at making bail decisions.
In my second book, Blink, I told the story of how orchestras made much smarter recruiting decisions once they had prospective hires audition behind a screen. Taking information away from the hiring committee made for better judgments. But that was because the information gleaned from watching someone play is largely irrelevant. If you’re judging whether someone is a good violin player, knowing whether that person is big or small, handsome or homely, white or black isn’t going to help. In fact, it will probably only introduce biases that will make your job even harder.
But when it comes to a bail decision, the extra information the judge has sounds like it should be really useful. In an earlier case in Solomon’s courtroom, a young man in basketball shorts and a gray T-shirt was charged with getting into a fight with someone, then buying a car with the man’s stolen credit card. In asking for bail, the district attorney pointed out that he had failed to appear for his court date after two previous arrests. That’s a serious red flag. But not all “FTAs” are identical. What if the defendant was given the wrong date? What if he would lose his job if he took off work that day, and decided it wasn’t worth it? What if his child was in the hospital? That’s what the defendant’s lawyer told the judge: Her client had a good excuse. The computer didn’t know that, but the judge did. How could that not help?
In a similar vein, Solomon said the thing he’s most alert to in bail cases is “mental illness with an allegation of violence.” Those kinds of cases are a judge’s worst nightmare. They let someone out on bail, then that person stops taking their medication and goes on to commit some horrible crime. “It’s shoot a cop,” Solomon said.
It’s drive a car into a minivan, killing a pregnant woman and her husband. It’s hurt a child. [It’s] shoving somebody in front of a subway train and killing them. It’s an awful situation at every possible angle.…No judge would ever want to be the one having made the release decision on that case.
Some of the clues to that kind of situation are in the defendant’s file: medical records, previous hospitalizations, some mention of the defendant’s being found not competent. But other clues are found only in the moment.
“You also will hear terms thrown around in the courtroom of ‘EDP’—emotionally disturbed person,” Solomon said.
That will come from either the police department who’s brought them in and handed you an envelope that’s from a doctor at a hospital where he’s been screened at a psychiatric ER prior to arraignment.…Other times, that information will get into the DA’s folder and the DA will ask questions.…That’s a fact for me to think about.
He’ll look at the defendant, in those cases—closely, carefully, searching for, as he put it,
sort of a glassy-eyed look, not being able to make eye contact. And not the adolescent unable to make eye contact because the frontal lobe hasn’t developed. I’m talking about the adult off their meds.…
Mullainathan’s machine can’t overhear the prosecutor talking about an EDP, and it can’t see that telltale glassy-eyed look. That fact should translate into a big advantage for Solomon and his fellow judges. But for some reason it doesn’t.
Puzzle Number Two: How is it that meeting a stranger can sometimes make us worse at making sense of that person than not meeting them?
5.
Neville Chamberlain made his third and final visit to Germany at the end of September 1938, two weeks after his first visit. The meeting was in Munich at the Nazi Party’s offices—the Führerbau. Italian leader Benito Mussolini and French prime minister Édouard Daladier were also invited. The four of them met, with their aides, in Hitler’s private study. On the morning of the second day, Chamberlain asked Hitler if the two of them could meet alone. By this point, Chamberlain felt he had the measure of his adversary.
When Hitler had said his ambitions were limited to Czechoslovakia, Chamberlain believed that “Herr Hitler was telling the truth.” It was now just a matter of getting that commitment in writing.
Hitler took him to his apartment on Prinzregentenplatz. Chamberlain pulled out a piece of paper on which he had written a simple agreement and asked Hitler whether he would sign it. As the interpreter translated the words into German, “Hitler frequently ejaculated, ‘Ja! Ja!’ And at the end he said, ‘Yes I will certainly sign it,’” Chamberlain later wrote to one of his sisters. “‘When shall we do it?’ I said, ‘now,’ & we went at once to the writing table & put our signatures to the two copies which I had brought with me.”
That afternoon, Chamberlain flew home to a hero’s welcome. A crowd of journalists surged toward him. He took the letter from his breast pocket and waved it to the crowd. “This morning I had another talk with the German Chancellor Herr Hitler, and here is a paper which bears his name upon it as well as mine.”
Then it was back to the prime minister’s residence at 10 Downing Street.
“My good friends, this is the second time in our history that there has come back from Germany to Downing Street peace with honor. I believe it is peace for our time. We thank you from the bottom of our hearts.”
The crowd cheered.
“Now I recommend you go home, and sleep quietly in your beds.”
In March 1939, Hitler invaded the rest of Czechoslovakia. It had taken him less than six months to break his agreement with Chamberlain. On September 1, 1939, Hitler invaded Poland, and the world was at war.
We have, in other words, CIA officers who cannot make sense of their spies, judges who cannot make sense of their defendants, and prime ministers who cannot make sense of their adversaries. We have people struggling with their first impressions of a stranger. We have people struggling when they have months to understand a stranger. We have people struggling when they meet with someone only once, and people struggling when they return to the stranger again and again. They struggle with assessing a stranger’s honesty. They struggle with a stranger’s character. They struggle with a stranger’s intent.
It’s a mess.
6.
One last thing:
Take a look at the following word, and fill in the two blank letters. Do it quickly, without thinking.
G L _ _
This is called a word-completion task. Psychologists commonly use it to test things such as memory.
I completed G L _ _ as GLUM. Remember that. The next
word is:
_ _TER
I completed that as HATER. Remember that too. Here are the rest of the words:
S_ _RE
P_ _ N
TOU_ _
ATT_ _ _
BO_ _
FL_ _ T
SL_ T
STR_ _ _
GO_ _
CHE_ _
_ _OR
SL_ _ _
SC _ _ _
_ _ NNER
B_ _ T
PO _ _ _
BA_ _
_RA_
_ _ _EAT
I started out with GLUM and HATER and ended up with SCARE, ATTACK, BORE, FLOUT, SLIT, CHEAT, TRAP, and DEFEAT. That’s a pretty morbid and melancholy list. But I don’t think that says anything about the darkness of my soul. I’m not melancholy. I’m an optimist. I think that the first word, GLUM, popped into my head, and then I just continued in that vein.
A few years ago, a team of psychologists led by Emily Pronin gave a group of people that same exercise. Pronin had them fill in the blank spaces. Then she asked them the same question: What do you think your choices say about you? For instance, if you completed TOU_ _ as TOUCH, does that suggest that you are a different kind of person than if you completed it as TOUGH? The respondents took the same position I did. They’re just words.
“I don’t agree with these word-stem completions as a measure of my personality,” one of Pronin’s subjects wrote. And the others in the group agreed:
“These word completions don’t seem to reveal much about me at all.…Random completions.”
“Some of the words I wrote seem to be the antithesis of how I view the world. For instance, I hope that I am not always concerned about being STRONG, the BEST, or a WINNER.”
“I don’t really think that my word completions reveal that much about me.… Occurred as a result of happenstance.”
“Not a whole lot.… They reveal vocabulary.”
“I really don’t think there was any relationship.… The words are just random.”
“The words PAIN, ATTACK, and THREAT seem similar, but I don’t know that they say anything about me.”
But then things got interesting. Pronin gave the group other people’s words. These were perfect strangers. She asked the same question. What do you think this stranger’s choices reveal? And this time Pronin’s panel completely changed their minds.
“He doesn’t seem to read too much, since the natural (to me) completion of B_ _K would be BOOK. BEAK seems rather random, and might indicate deliberate unfocus of mind.”
“I get the feeling that whoever did this is pretty vain, but basically a nice guy.”
Keep in mind that these are the exact same people who just moments before had denied that the exercise had any meaning at all.
“The person seems goal-oriented and thinks about competitive settings.”
“I have a feeling that the individual in question may be tired very often in his or her life. In addition, I think that he or she might be interested in having close personal interactions with someone of the opposite sex. The person may also enjoy playing games.”
The same person who said, “These word completions don’t seem to reveal much about me at all” turned around and said, of a perfect stranger:
“I think this girl is on her period.…I also think that she either feels she or someone else is in a dishonest sexual relationship, according to the words WHORE, SLOT (similar to slut), CHEAT.”
The answers go on and on like this. And no one seemed even remotely aware that they had been trapped in a contradiction.
“I guess there is some relationship.…He talks a lot about money and the BANK. A lot more correlation here.”
“He seems to focus on competition and winning. This person could be an athlete or someone who is very competitive.”
“It seems this individual has a generally positive outlook toward the things he endeavors. Most words, such as WINNER, SCORE, GOAL, indicate some sort of competitiveness, which combined with the jargon, indicate that he has some athletic competitive nature.”
If the panel had seen my GLUM, HATER, SCARE, ATTACK, BORE, FLOUT, SLIT, CHEAT, TRAP, and DEFEAT, they would have worried for my soul.
Pronin calls this phenomenon the “illusion of asymmetric insight.” She writes:
The conviction that we know others better than they know us—and that we may have insights about them they lack (but not vice versa)—leads us to talk when we would do well to listen and to be less patient than we ought to be when others express the conviction that they are the ones who are being misunderstood or judged unfairly.
This is the problem at the heart of those first two puzzles. The officers on the Cuba desk of the CIA were sure they could evaluate the loyalty of their spies. Judges don’t throw up their hands at the prospect of assessing the character of defendants. They give themselves a minute or two, then authoritatively pass judgment. Neville Chamberlain never questioned the wisdom of his bold plan to avert war. If Hitler’s intentions were unclear, it was his job, as prime minister, to go to Germany and figure them out.
We think we can easily see into the hearts of others based on the flimsiest of clues. We jump at the chance to judge strangers. We would never do that to ourselves, of course. We are nuanced and complex and enigmatic. But the stranger is easy.
If I can convince you of one thing in this book, let it be this: Strangers are not easy.
1 The one exception was Canadian prime minister William Lyon Mackenzie King. He met Hitler in 1937. He loved him. He compared him to Joan of Arc.
2 The Nazi official Henderson knew even better was Göring, Hitler’s deputy. Henderson would go stag hunting with Göring. They had long conversations. Henderson was convinced that Göring wanted peace as well, and that underneath his Nazi bluster was a decent man. In a memoir of his time in Berlin, written just as war broke out, Henderson said that Göring “loved animals and children; and, before ever he had one of his own, the top floor at Karinhall contained a vast playroom fitted up with every mechanical toy dear to the heart of a modern child. Nothing used to give him greater pleasure than to go and play there with them. The toys might, it is true, include models of airplanes dropping heavy bombs which exploded on defenseless towns or villages; but, as he observed when I reproached him on the subject, it was not part of the Nazi conception of life to be excessively civilized or to teach squeamishness to the young.” (In case you were wondering, that’s what Nazism was really about: tough-minded child-rearing.)
3 The law has since been changed. A defendant must be eighteen years old or above to be sent to Rikers.
4 Two technical points about the dueling lists of 400,000 defendants: When Mullainathan says that the computer’s list committed 25 percent fewer crimes than the judge’s list, he’s counting failure to appear for a trial date as a crime. Second, I’m sure you are wondering how Mullainathan could calculate, with such certainty, who would or wouldn’t end up committing a crime while out on pretrial release. It’s not because he has a crystal ball. It’s an estimate made on the basis of a highly sophisticated statistical analysis. Here’s the short version. Judges in New York City take turns doing bail hearings. Defendants are, essentially, randomly assigned to them for consideration. Judges in New York (as in all jurisdictions) vary dramatically in how likely they are to release someone, or how prohibitively high they set bail. Some judges are very permissive. Others are strict. So imagine that one set of strict judges sees 1,000 defendants and releases 25 percent of them. Another set of permissive judges sees 1,000 defendants, who are in every way equivalent to the other 1,000, and releases 75 percent of them. By comparing the crime rates of the released defendants in each group, you can get a sense of how many harmless people the strict judges jailed, and how many dangerous people the permissive judges set free. That estimate, in turn, can be applied to the machine’s predictions. When it passes judgment on its own 1,000 defendants, how much better is it than the strict judges on the one hand, and the
permissive judges on the other? This sounds highly complicated, and it is. But it’s a well-established methodology. For a more complete explanation, I encourage you to read Mullainathan’s paper.
Part Two
Default to Truth
Chapter Three
The Queen of Cuba
1.
Let’s take a look at another Cuban spy story.
In the early 1990s, thousands of Cubans began to flee the regime of Fidel Castro. They cobbled together crude boats—made of inner tubes and metal drums and wooden doors and any number of other stray parts—and set out on a desperate voyage across the ninety miles of the Florida Straits to the United States. By one estimate, as many as 24,000 people died attempting the journey. It was a human-rights disaster. In response, a group of Cuban emigrés in Miami founded Hermanos al Rescate—Brothers to the Rescue. They put together a makeshift air force of single-engine Cessna Skymasters and took to the skies over the Florida Straits, searching for refugees from the air and radioing their coordinates to the Coast Guard. Hermanos al Rescate saved thousands of lives. They became heroes.
As time passed, the emigrés grew more ambitious. They began flying into Cuban airspace, dropping leaflets on Havana urging the Cuban people to rise up against Castro’s regime. The Cuban government, already embarrassed by the flight of refugees, was outraged. Tensions rose, coming to a head on February 24, 1996. That afternoon three Hermanos al Rescate planes took off for the Florida Straits. As they neared the Cuban coastline, two Cuban Air Force MiG fighter jets shot two of the planes out of the sky, killing all four people aboard.
Talking to Strangers Page 4