When our brain takes an educated guess

Last week, the Solicitor-General released the report into the case of Alan Hall. Since the Supreme Court quashed Hall’s conviction last year, people have referred to him as 'wrongfully convicted' or 'exonerated'. I’m not so sure.

There are two types of miscarriages of justice: wrongful convictions and procedural miscarriages. A wrongful conviction occurs when an innocent person is convicted of a crime they didn’t commit. This is known as 'actual innocence'.

A procedural miscarriage occurs when there was a significant breakdown in a procedural safeguard in the system – when police or prosecutors commit misconduct, defence counsel fails to provide effective representation, unreliable or unfair evidence is presented, or jurors are misdirected or misunderstand the meaning of proof beyond a reasonable doubt. These safeguards exist in part to prevent the innocent from being convicted, but they can fail in the trials of guilty people, too.

Teina Pora was wrongfully convicted. The semen found inside Susan Burdett came from Malcolm Rewa, a serial rapist with a lengthy history of crimes of sexual violence. Pora was convicted because of a 'confession' he gave after a police interrogation that was impermissibly suggestive, particularly given Pora’s neurodevelopmental and intellectual disabilities. He suffered a procedural miscarriage of justice (unfair police questioning) but was also actually innocent.

Hall’s conviction was quashed because his trial was fundamentally unfair. The Crown withheld – and police possibly tampered with – potentially exculpatory evidence relating to an eyewitness’s description of the likely perpetrator. Ronald Turner told police he was “100 percent sure” the person he saw was Māori, and Hall is Pākehā. At Hall’s preliminary hearing, Turner’s witness statement had any reference to the fleeing suspect as Māori removed. The police officer or prosecutor who produced Turner’s misleading statement violated Hall’s right to a fair trial and possibly committed obstruction of justice.

None of this means Hall is innocent. Easton was killed with Hall’s bayonet. Hall has never given a convincing or consistent explanation of how his bayonet became the murder weapon. He had no alibi, and he was in the vicinity of the Easton home at the time. He absolutely deserved to be tried with all relevant evidence relating to his guilt or innocence being presented to the jury, but he has not been conclusively 'exonerated'.

Turner’s statements about the suspect he saw are also worrying. Yes, he originally told police he was sure the person was Māori. But he also told police he didn’t get a good look at the man because it was dark, his face was covered by a hoodie, and he didn’t see his facial features. His conclusion the fleeing suspect was Māori appears to be based primarily on his assessment that the man was “definitely dark skinned, not white”.

Turner’s statements remind me of the 'gorilla studies' – psychological studies about the way human beings selectively and subconsciously process information. In the original gorilla study, test subjects were shown a video of a multicultural group of children playing basketball. The subjects were told they were participating in a study about subconscious biases (true). They were supposed to watch the game and count how many times the children passed to their white and black team mates, respectively. In the middle of the video, a man in a gorilla suit walked straight through the game. The test subjects did an excellent job counting the bounce passes, but more than half failed to notice the gorilla. This launched a series of follow-up studies, all of which found some variation on the same theme: humans are really bad at noticing gorillas when we’re focused on something else.

In one study, researchers had radiologists look at mammogram films. They told them they were measuring their accuracy in identifying breast cancer (false). They showed them six mammograms, three with tumours and three without. Breast tumours are incredibly small, and reading mammograms is a highly specialised skill. If you’re not a radiologist, you probably couldn’t see breast cancer on a mammogram if they pointed it out to you. The study produced some reassuring news: the radiologists detected the tumours about 99 percent of the time. It also produced some less-than-reassuring gorilla-related news. On all six films, researchers had digitally added a picture of a gorilla. The gorilla was about an inch squared – much larger than a typical cancer nodule. Less than 20 percent of the radiologists noticed the gorilla. When researchers later pointed out the gorilla, most insisted it hadn’t been there when they reviewed the film (false).

This phenomenon is known as 'inattentional blindness', and it’s not always bad. Inattentional blindness is what allows you to drive a car through the streets, noticing traffic lights and other cars but tuning out open homes and billboards. But it can also distort perception.

What does this have to do with Turner’s eyewitness evidence? I have no doubt Turner believed he saw a Māori man fleeing the scene. I am less convinced his belief is reliable.

Studies about stereotyping demonstrate we all have subconscious associations between people with particular features (Māori men) and characteristics our culture associates with those features (criminality, dangerousness). These implicit associations have no relationship to people’s philosophical beliefs about racial equality. White supremacists are no more likely to have them than civil-rights leaders.

Studies about human perception also show the mind sometimes works a bit like a digital camera. The reason digital cameras produce pictures with higher resolution than film is their software fills in missing information. It takes the actual pixels of the photograph and makes educated guesses about what lies between to create clearer, fuller images. The human brain sometimes does the same thing when we have partial perceptual information. It takes an educated guess, based on our lived experiences (and unconscious stereotypes) and fills in the blanks. Your brain then comes to believe the resulting filled-in picture is your original memory. This is called the 'relation-back' phenomenon.

When you combine inattentional blindness, implicit associations, and relation-back, they shed new light on Turner’s description of what happened the night of Easton’s murder. By his telling, he looked out his window and saw a man in dark clothes, half walking, half running, while looking back over his shoulder. In other words, he saw someone suspicious. He didn’t get a good look at the guy, but he thought his skin looked dark. Given Turner is human, his brain may have taken the information it had (shady, suspicious, dark-looking), filled in the information it didn’t have (facial features, unobstructed view, good lighting), and come up with 'Māori'.

I’m not saying Turner didn’t see a Māori man running away from the Easton house or that he really saw Alan Hall. I’m certainly not suggesting Turner is racist. Not at all. He’s just human and, like the rest of us, lives in a culture with deeply-embedded stereotypes. His focus on the suspect’s suspicious behaviour may have prevented him from noticing the gorilla (the suspect wasn’t Māori).

Police and prosecutors have trouble spotting gorillas, too. Once they’ve locked onto a suspect and are convinced of that person’s guilt, they can become inattentionally blind to exculpatory information that suggests the person is innocent.

I think police and prosecutors could have believed Turner was mistaken about what he saw. If so, there was a wrong way to handle it (editing the damaging, unreliable information out of his statement and not telling the defence) and a right way to handle it. The right way would have involved putting all the information before the jury so they could do their constitutional job. It might even have involved calling a cognitive psychologist to offer context evidence about processes like implicit associations and confirmation bias. The jury might have retired to the deliberation room and decided, “You know what? I know he said he was sure the guy was Māori, but I just don’t think we can rely on that anymore.” They might have retired to the jury room and said, “He said he was sure the guy was Māori. He might have been wrong, but we can’t be sure enough to be convinced of guilt beyond a reasonable doubt.” Either way, the system would have worked the way it’s supposed to.