Selective Perception


Tim Couch was the quarterback taken first overall in the 1999 NFL Draft. Couch has a baby’s blue eyes and a chin that tapers down to a point like a guitar pick. As a college sophomore, Couch had turned what had been a pathetic Wildcats offense into the leading unit in the nation, setting several new school records along the way. In 1998, he was a Heisman finalist and named SEC Player of the Year. Couch had the prototypical measurables as well, standing 6’4 and weighing 225 lbs. One NFL scout, Don Shonka, described Couch’s accuracy like this: “They used to put five garbage cans on the field and Couch would stand there and throw and just drop the ball into every one.” Couch never lived up to his billing in the pros. After just five years in the league, appearing in 62 games, Couch was cut by the Browns and didn’t play again. Being able to drop the ball into garbage cans, as impressive as that is, does precious little to predict how well you can throw a ball to a sprinting receiver.

It is notoriously difficult to identify what college players will turn into successful professional players. When it comes to quarterbacks, this statement is especially true. Tim Couch hardly stands alone amongst promising – record-setting, even – collegians who would be called “busts.” There’s Joey Harrington, sarcastically called “Joey Blue-Skies” by the Detroit press for his predictably optimistic post-game interviews. There’s Akili Smith, who was drafted 3rd overall by the Bengals and would complete less than half of his passes as a pro. 2007’s number one overall pick was Jamarcus Russell, whose contract with Oakland guaranteed him $31.5 million. The Raiders’ ROI was 15 lost fumbles. Russell’s competition for biggest draft bust of all time was the Charger’s Ryan Leaf, whose rookie season featured 2 touchdown passes and 15 interceptions. Leaf is currently in prison for felony burglary.

Imagine you are asked to look at five college quarterbacks and asked the question, “Which one of these is most likely to be a good NFL quarterback?” How would you approach this problem? You might sort them by their physical characteristics, such as height, weight, hand size, or arm length. You could measure how far they could throw the ball, and how accurate those throws were, and how fast they can run. You might want to test their peripheral vision, and their intelligence. You would probably give some weight to their college statistics and what type of program they played in. You could ask their teammates what kind of work ethic each player displayed, or how much passion they have. This is the approach NFL teams take. They spend millions of dollars and countless hours doing exactly these things. And despite having every metric and measurement imaginable, they are little better than casual fans at predicting who will be succeed.

This is the scouting problem: there are certain roles where almost nothing you learn about a candidate beforehand can predict how well they’ll perform. But you don’t need to be an NFL scout to face it. The scouting problem shows up when schools are hiring teachers or firms are hiring financial advisers. Finding the right person can be an enormously complicated question with no clear answer.


Just after midnight on the morning of November 28th, 1976, Officer Robert Wood pulled over a car that was driving with its headlights out. When he approached the car, Wood was shot and killed. (In a tragic side note, bulletproof vests were apparently not yet standard issue equipment for the Dallas Police Department. Wood’s wife had purchased one for him and was giving it to him as a Christmas present.) A month passed before police got a lead in the case: a sixteen-year old kid named David Harris had been bragging to his friends that he killed a cop. After interviewing Harris, the police found 1) the vehicle driven by the suspect had been stolen from Harris’ neighbor; 2) ballistics from the gun used to kill Officer Wood matched the gun Harris had stolen from his father, a .22 revolver; 3) this gun was still in Harris’ possession; 4) Harris matched the description provided by Wood’s partner; and, 5) between the murder and the interview, Harris had been arrested for holding up a convenience store. Not surprisingly, Harris denied shooting Officer Wood.

What is surprising, however, is the fact that the investigating officers believed him. Based only on Harris’ testimony, police arrested an Ohio man named Randall Adams. Adams had met Harris the day before the shooting, when Harris offered him a ride. The two spent much of the day together, drinking and smoking marijuana. Adams claims he was back at his motel at the time of the shooting. Bill James, author of Popular Crime, put it this way. “It was Harris who had shot the cop. In retrospect, there is every reason why the police should have known this. Adams, 28 years old at the time, had never been in any trouble. Harris had been involved in criminal activity all his life…. It is anyone’s guess why the police and prosecutors chose to believe the wrong man.” But believe them they did. Maybe it was, as James puts it, because “Harris was a personable young man.” Maybe it was because the prosecutor wanted to send someone to the electric chair for killing a cop, and thus chose to charge the adult rather than the minor.

We expect the police to be fundamentally incapable of making this kind of error. They should be objective, we might say. They should focus on the facts. But step back a moment and place yourself in the detective’s shoes. A cop was murdered, shot through the heart. Maybe you knew him and he was in your regular poker game. Or maybe he was from a different precinct – Dallas is not a small town – but just the notion of an officer being gunned down would likely be enough to unsettle your focus a little bit. On the one hand, you have the likable, gregarious kid who’s telling you it wasn’t him …and who would be facing a short prison sentence at most. On the other hand, you have an outsider. Someone old enough to pay with his life. So what if the evidence all points to the other guy? With a little distance and the benefit of hindsight, it’s easy to see how one could get tunnel vision.


How can we predict who will make a good spouse? In a perverse sort of way, marriage is the combination of both of these problems. On the one hand, we have a scouting problem. When we are dating someone – the college athletics of relationships – we accumulate all kinds of data, whether these are mere impressions or things we can put into quantifiable terms. The problem isn’t having enough information: we have too much information. The problem is knowing where to place our focus. We can’t even say which factors best predict success in the NFL; how are we supposed to know what characteristics are most important to look for in a spouse? On the other hand, we can get so lost in the emotional thicket of a romance that it’s easy to develop romantic tunnel vision. Obvious signs like, say, drug use, or prior criminal behavior, or patterns of aggression get ignored because he’s just oh-so-likable and gregarious. Even when the red flags aren’t quite so glaring, we might overlook a lack of compatibility because of a strong physical attraction; conversely, we might fail to notice deep compatibility when that attraction is less intense.

John Gottman is a psychologist who is best known for his ability to predict divorces: if you get a couple to discuss a contentious aspect of their relationship in front of Gottman for just five minutes, he can tell you with 91% accuracy whether or not they will be divorced within ten years. It’s fair to say that Gottman has some understanding of how relationships work. One thing he likes to talk about is active listening. Many marriage counselors recommend couples develop better active listening skills: feed back to your spouse what she’s saying to make it clear you understand and empathize. To Gottman, this is not a particularly useful habit. “Let’s say my wife is really angry with me because I repeatedly haven’t balanced the checkbook and the checks bounce. What would it accomplish if I say: ‘I hear what you’re saying, you’re really angry with me, and I can understand why you’re angry with me because I’m not balancing the checkbook.’ That’s not going to make her feel any better, I still haven’t balanced the damned checkbook!”

In Gottman’s eyes, real empathy is not found in an effective summary of the other person’s point of view. “Real empathy comes from going: ‘You know, I understand how upset you are. It really hurts me that I’m messing up this way, and I’ve got take some action.’ Real empathy comes from feeling your partner’s pain in a real way, and then doing something about it.” In a way, active listening is like being able to throw a football into a garbage can forty yards down the field: showing you can parrot a thought doesn’t show you’re capable of empathy. Gottman describes what he calls “relationship masters” as the sort of people who look at the problems in their relationship and develop a collaborative approach. “We’ve got this problem. Let’s take a look at it, let’s kick it around. How do you see it? I see it this way, and we kick it around,” he says. “All of a sudden I can have empathy for your position because you’re telling me what you contribute to the problem.”

Of course, there’s more to the story. The ability to succeed in a relationship isn’t all about one specific trait any more than playing quarterback is about the ability to take a shotgun snap. But the point isn’t to describe all the factors that make a successful relationships – I’m not qualified to do that. The point is that no matter what factors those are, the process has to start as an introspective one. “How am I contributing to our problems? What can I do to get better?” You cannot collaborate by shifting blame.


In a classic experiment, subjects were shown a sequence of playing cards, each for a fraction of a second. They were asked to identify the cards. For four of the cards, there was complete agreement among all subjects. Everyone accurately picked out the seven of diamonds, say, or the nine of clubs. But on one card, opinions were divided. Some people saw a three of hearts. Some thought it was a three of spades. The reality? It was a three of hearts, but the numbers and the shapes were colored black rather than red. Some people noticed the shape and filtered their perception through that expectation: it didn’t even occur to them that the hearts were colored black. Others noticed the color, which in turn forced them to conclude that the shapes were spades. When faced with conflicting stimuli and the pressure of a time constraint, one factor dominates our perception and the other is pushed out of the way entirely.

In psychology, these failures fall under the set of cognitive biases called selective perception. Selective perception is the process by which we base conclusions on a small amount of available information and ignore what contradicts it. According to Sarah Mae Sincero, it is the process “in which a person only perceives what he desires to and sets aside or ignores other perceptions or viewpoints.” (The poet Anais Nin put it like this: “We do not see things as they are. We see things as we are.”) Selective perception is why, when the Packers play the Bears, Bears fans are convinced the Packers committed more penalties. And selective perception is why the Dallas Police Department arrested Randall Adams for the murder of Robert Wood. Harris was a likable kid with all kinds of evidence pointed at him. He was a black three of hearts. For the police and prosecution, seeing the shape was all they needed to know about the color.

Every person on the planet commits selective perception errors. The consequences of these are typically small: you disagree with someone about a pass interference call. If you step back and ask yourself, even occasionally, “Am I motivated to see this a certain way? What expectations or beliefs did I have before I had any reason to believe them? Would I see things differently if I didn’t have those motivations?” you might be able to protect yourself from making these mistakes when the stakes are much higher, whether those stakes are a criminal prosecution or a relationship.


Russell Wilson was a bit of conundrum for NFL scouts. Some scouts thought he was among the best quarterback prospects to ever enter the NFL draft. He had excelled at two different colleges – something that is virtually unheard of. Scouts had trouble pinpointing any negatives in his game. But at the same time, he is short: Super Bowl winning coach Jon Gruden said, “The only negative is his height.” Most teams want a quarterback who is at least 6’4, someone who can see over lineman and identify receivers down the field. Wilson is 5’11. Scouts ultimately decided that Wilson’s height overwhelmed his other traits. Wilson fell to the third round of the draft, going 75th overall. Those scouts were wrong. If the 2012 NFL draft were done over today, Wilson wouldn’t go any later than the 3rd choice.

One big error scouts – and casual observers – make before the NFL draft is they fundamentally misunderstand how to interpret Combine data. (If you don’t know, the Combine is where all prospective draftees go to get measured, timed, and quizzed so we can know how long their arms are, how fast they can sprint 40 yards, and how many feet of rope you can buy for $.40.) Everyone ends up focusing on the biggest, strongest, and fastest. But that isn’t the point of the combine. The point of the combine is to determine whether someone meets the minimum baseline for NFL success. A wide receiver that runs a 4.5 second 40-yard dash is probably fast enough. One who runs a 6.5 second 40-yard dash is not. Because they were so focused on the fact that Russell Wilson wasn’t as tall as, say, Jamarcus Russell, they missed the fact that he was tall enough. He met the baseline to play his position.

Michael Robinson is Wilson’s fullback in Seattle. One night after practice, Wilson had Robinson stay to work on every last, excruciating detail of the most basic passing route in the tree: the flat route. Wilson was not content until he felt he could consistently throw to the point that would allow Robinson to turn upfield without losing momentum. Wilson’s father told him, “The separation’s in the preparation.” It’s an aphorism he passed onto Robinson. There are other reasons Russell Wilson is fast becoming one of the league’s better quarterbacks. He plays for a coach that tailors the offense to his strengths. He has an excellent corps of receivers and one of the best running backs in the league. Ultimately, though, he works until he gets it right. He asks himself how he can be better, and then goes and makes himself better. Russell Wilson should never had lasted to the 75th pick in the draft. But even with how closely scouts watched him, all they could see was his height. It’s almost like they said, “Yeah, yeah, yeah. He’s everything you’d want a quarterback to be. But did you see this guy? He’s six foot four.”


Addendum: Who Killed Mary Rogers?

“I’ve been listening to my gut since I was 14 years old, and frankly speaking, I’ve come to the conclusion that my guts have shit for brains.”
– Rob Gordon, High Fidelity

Ever work on a puzzle, or a riddle, or a brain teaser that just completely stumped you in the moment? Assuming the answer is yes, ever have one that had such a clear, obvious solution that you berated yourself for missing it? One of the unfortunate realities of crime is the fact that there are so many unsolved killings and disappearances. Some of these are baffling and mysterious. Others, though, ultimately have obvious solutions. I believe the case of Mary Rogers has an obvious solution. In my previous post, I dismissed one of the leading explanations given at the time: that Mary died as a result of a botched abortion. Let me overview some competing explanations before I explain who I think is responsible in this 170-year old cold case.

1) Either Daniel Payne or one of Mary’s numerous romantic entanglements killed her. Witnesses placed Mary in Hoboken, New Jersey, with a tall, dark-complexioned gentleman on the day of her death. Mary’s itinerary was to go to Nassau St. in Manhattan. It is entirely plausible that this man murdered Mary and fled the region. It is also plausible that Daniel Payne, her boss John Anderson, or ex-boyfriend Arthur Crommelin followed Mary to New Jersey and, jealous of the new suitor, murdered them both. No witnesses could place either of the latter three near the scene of the crime, Payne had an alibi, and despite intense focus from both police and press, none of these three men were ever linked to the crime in any meaningful way.

2) A New York street gang murdered Mary and her companion. Dr. Richard Cook, the coroner who performed Mary’s autopsy, says he believed she was attacked by between six and eight men. It’s unclear to me how he could make such a claim. I would be more inclined to believe a less-specific statement: “Mary was attacked by multiple men.” No resource I could find on the case explained Cook’s thought process. Frederica Loss, the proprietor of Nick Moore’s House – the last place Mary was seen alive – initially also advanced this theory, saying she heard a scream after Mary left the tavern. Her sons also reportedly found several of Mary’s garments, including a monogrammed handkerchief. I consider this explanation both unsatisfactory and implausible.

3) Mary Rogers died during a botched abortion. This story originates from Frederica Loss. She said, in dramatic deathbed fashion, that Mary came with the mysterious dark-complexioned man, to her tavern to have an abortion. The procedure went awry, and Mary bled to death. Her companion disappeared from the scene, leaving Loss to handle the dead body. She instructed her sons, aged 16, 18, and 20, to throw it in the Hudson River. The New York media jumped on this version of events, leading abortion to become outlawed in New York within three years. It doesn’t match Dr. Cook’s autopsy, however: Mary had been bound, beaten, and strangled. Her wounds were consistent with rape, not an abortion.

Easily the most plausible explanation I have read is also the simplest and most straightforward. Frederica Loss was one of the last persons to see Mary alive. Witnesses placed Mary at her tavern. She changed her story about that encounter at least once. She admitted her sons disposed of Mary’s body. In her original version of events, she claims her sons “found” Mary’s clothing. The balance of probability suggests that Mary Rogers was raped and murdered by Frederica Loss’ sons and she attempted to cover up the crime. This version of events was not considered for another 65 years, when a criminologist named Bill Clemens investigated the case for Era Magazine. As I read through the account the first time, my eyebrow raised at the fact that the Loss boys were never put under suspicion. When one of them “discovered” Mary’s garments, I said aloud, “Wait, how old were these boys?” To be fair, I’ve seen at least three episodes of Law & Order: SVU. I’m practically old hat by now.


The Mary Rogers murder is discussed at some length in Bill James’ excellent book Popular Crime. James advocates the solution I present. A running theme in Popular Crime is how investigators can get so convinced their suspect is the guy that they try to shoehorn evidence around that person. What they should be doing is paying attention to who is already linked to the evidence. (That’s easier said than done, though. Rarely is anything simple about a murder investigation.) I’m talking about this case because I think we are constantly conducting our own investigations. And, like the police in New York and Hoboken, we make a lot of ridiculous errors that become obvious in retrospect. We have questions about who to trust, or who to date, or who to vote for, or what is so-and-so up to? Life likes to leave us evidence. Sometimes we think we’ve got our guy, when a much more obvious candidate has been there all along.

Chris Kluwe and the Power of Storytelling

“Artists use lies to tell the truth. Yes, I created a lie. But because you believed it, you found something true about yourself.”
― Alan Moore, V for Vendetta

Mary Rogers worked in a cigar shop in New York City. Brown-eyed and brunette, her beauty was considered legendary. Patrons would spend all afternoon in the cigar shop, stealing glances and, importantly, smoking cigars. Another admirer even went so far as to publish a poem in the New York Herald, extoling her “heaven-like smile” and “star-like eyes.” Rogers was engaged to an alcoholic cork cutter named Daniel Payne. One morning in July of 1841, she told Payne and her mother she was going to visit her aunt. Rogers’ body was found floating in the Hudson River three days later. Public suspicion fell largely on Payne until a tavern owner named Frederica Loss “confessed” that Rogers had died during a botched abortion in Loss’ Hoboken tavern. This account almost certainly wasn’t true – the coroner reported that Rogers had been raped repeatedly and then strangled, and found no evidence of a pregnancy. Pesky little things like facts hardly matter in the maelstrom of public opinion: abortion in New York was legal in 1841; by 1845, this had changed.
Mary Rogers
In major crime stories, truth plays second fiddle to narrative. (That may be an exaggeration. It’s probably more accurate to say that truth was denied an audition due to its uncomely appearance.) Currents of storytelling, emotional themes, these are what resonate with the public at large. It is the narrative – distilled and dehydrated for mass shipping, like concentrated orange juice – that determines our emotional involvement in these issues. And this is fine for water cooler conversations and Facebook rants. But the Mary Rogers case isn’t an isolated incident: public understanding (or misunderstanding) of a major crime leads to legislative changes. School shootings pressure lawmakers to limit the types of cosmetic features the average citizen can purchase on their guns. This is an entirely superficial law: it shouldn’t make you feel any safer knowing your attacker’s shotgun doesn’t have a pistol grip. The Trayvon Martin shooting has led to a petition – with over 400,000 signatures – demanding an end to Stand Your Ground laws. The problem? Stand Your Ground had no bearing on the Trayvon Martin shooting, was not a component George Zimmerman’s defense, and had those laws never existed, Trayvon Martin would still be dead and George Zimmerman would still have been acquitted.

Last Thursday, former Minnesota Vikings punter Chris Kluwe published a scathing piece on Deadspin, claiming his contract was terminated by the Vikings due to his political activism. In it, he alleges that his position coach Mike Priefer made homophobic and bigoted remarks to mock Kluwe’s support for gay marriage. And already, before any facts have come out to confirm or disconfirm Kluwe’s statements, people are taking sides and signing petitions. (It probably doesn’t help matters that this came so fast on the heels of the Jonathan Martin and Phil Robertson debacles.) Let’s review what is actually known at this point.

1) Kluwe claims that Priefer made several homophobic remarks. Disgruntled ex-employees have been known to exaggerate.
2) Priefer denies saying these things. Accused persons, both guilty and not guilty, have been known to make denials.
3) Vikings kicker Blair Walsh, someone who would absolutely have been witness to Priefer’s remarks denies that they happened. Current employees often have the sense to not bite the hand that feeds them.
4) After being cut by the Vikings, Kluwe was unable to find further work in the NFL. Aging specialists with declining statistics are often replaced by younger, cheaper players.

All this is to say, we know very little about this case. Everyone who has made a statement thus far has a clear motive to say exactly he said.

Priefer Kluwe

It is unclear what will happen to Priefer from here. The Vikings have hired investigators, though that has the feel of a PR move. There will probably be a lawsuit or two. It seems likely that the NFL will make some rule changes about hazing and harassment. Since the NFL is influential, many of these changes will trickle down into college and high school organizations. Maybe those changes will make it possible for harassed players to protest their treatment without the fear of losing their jobs or future opportunities. Those would be good things. But the fact of the matter is, the facts here don’t really matter. It doesn’t matter if Priefer said what Kluwe claims. It makes for an emotionally-compelling story. And storytelling, not facts, drive change.