‘Our Lying Eyes’: Inaccurate eyewitness identifications and wrongful convictions | Hacker News Reader

‘Our Lying Eyes’: Inaccurate eyewitness identifications and wrongful convictions

nybooks.com

74 pointshhs36 comments

I was asked to provide an eyewitness identification for a crime I witnessed being committed and reported to the police. I had about a minute-long conversation with one of the perpetrators.

When the detective came to my home with a black and white photo lineup, she had a series of photos of people who looked very similar to one another. She explained that lineups were intentionally set up with people who look alike, to ensure that I was dead sure of who I was identifying. Long story short, I declined to make an identification because even after talking to someone and looking straight at him for a full minute, I wasn’t sure enough to become an eyewitness and potentially be wrong about it.

I recently went to a presentation by the Arizona Justice Project [0], an organization that works to free the wrongfully convicted. They said something about how the best way to do a lineup is to present photos one at a time, with no prompting whatsoever.

The Department of Justice has various pages about lineups, this one seems informative: https://www.justice.gov/opa/pr/justice-department-announces-...

[0] http://www.azjusticeproject.org/

Psychology videos of change blindness and inattention blindness are fun. For instance, Simons and Levin's "door" study swaps conversation partners in the midst of their getting directions. Swaps of similar people are less likely to be noticed, but sometimes even very dissimilar (different gender and race and height) swaps aren't. Though to be fair, there's often a misdirection task and/or object.

This evening there was someone seemingly "stalking" around on the subway. Perhaps plainclothes, or crook, or artist, or just someone a bit unusual. Getting off, and pondering the ethics of mentioning it, I realized it was a non issue, as I couldn't give a useful description of someone I'd been puzzled by for several minutes. Later, walking down the street, looking at someone, looking away and generating a description, and then looking back to check it... it's clearly not a skill I've developed.

It's worth noting that almost all photo lineups include only the suspect and a bunch of photos they keep around. They generally aren't asking you to choose between a group of possible suspects; they're using the other photos to make sure that you're positive. You should try to make an effort, because even an incorrect identification provides some amount of deniability for the suspect.

No you absolutely should not. If you aren't sure, say so. Best guess is not the way to do this. Fingering their one suspect among a bunch of photos they keep around who looks maybe a bit like the person you saw? Happens all the time, innocent people go to jail. And yeah, especially if they're from a race with homogenous hair colour and that race is one you don't share.

Every time they have the wrong suspect. There are 10 photos. 1 in 10 chance of fingering that wrong suspect and providing false evidence contributing to wrongful prosecution.

I actually don't understand why you would even suggest that this is a good idea.

Is there something here I'm missing?

You should say you're not sure obviously, but tell them that you think it might be person 3.

Additionally, police almost always do a second and sometimes third lineup with other pictures if there's a positive identification.

> Finally, in 1993 [nine years later], DNA analysis of the semen extracted from the girl’s underwear showed that the culprit was not Bloodsworth but someone else, who eventually confessed. Thankfully, Bloodsworth had not been executed, and he was set free that year.

> Over a decade later, DNA testing of the semen taken from her vagina right after the crime proved to be that of another man, who was then charged, and Cotton was set free.

> Finally, however, after Bain obtained assistance from a lawyer and the Innocence Project of Florida, DNA testing was granted—and it completely exonerated him. He was released in 2009, having served thirty-five years for a crime he did not commit.

There ought to be a federal standard for compensating people who have been wrongly imprisoned.

> For example, Texas compensates the wrongfully convicted $80,000 per year of incarceration and an annuity set at the same amount ...

That is, you get $80K per year for the rest of your life. That seems pretty fair. But there also ought to be something like a life insurance payout for families of prisoners who were wrongly executed.

> But there also ought to be something like a life insurance payout for families of prisoners who were wrongly executed.

Or alternatively, have the state stop killing people.

Well, yes. That too.

pretty fair? you would trade your youth for $80k a year? i think that's an egregiously low amount for the way a person in prison spends that time.

> One other modest mitigating factor should be mentioned. For many everyday crimes, like robbery, the presence of surveillance cameras in stores and buildings has made the police somewhat less dependent on eyewitness identification. The broader use of such surveillance cameras should therefore be encouraged.

Oy vey, the cure may be worse than the disease.

We’ve know this for over a century[0]. Plane crash investigators mostly ignore eye witness testimony using it only at best to guide the investigation[1]. Maybe this should be the standard for the legal system too: allow eye witness testimony to guide police investigation but not be used as courtroom evidence.

Also, this article doesn’t mention the cross or other race effect[2] which is yet another compounding factor.

Aside: Rashomon is still an excellent movie.

0. http://historyforensicpsych.umwblogs.org/eye-witness-account...

(The link above embeds a 60 minutes segment about the Ronald Cotton case mention in the submitted story.)

1. https://commons.erau.edu/cgi/viewcontent.cgi?article=1040&co...

2. https://www.psychologytoday.com/us/blog/culture-conscious/20...

  Maybe this should be the standard for the legal system 
  too: allow eye witness testimony to guide police
  investigation but not be used as courtroom evidence.

I used to think this, until I had to sit on a jury. I ended up trying a case where the accused had allegedly gone to the victim's home, knocked on the door, then barged in and punched the victim a few times when he answered the door. Hospital records and photos meant no-one disputed the victim had been punched. But what evidence was there that the accused was the attacker? Only the eyewitness testimony of the accused and his wife.

I struggle to see how such a case could be prosecuted without eyewitness testimony? Unless perhaps people put CCTV cameras all around in their own homes, and to me that doesn't sound like an improvement in civil liberties.

(After we found the guy guilty, the judge revealed the accused was awaiting trial for two other very similar offences with different victims)

>I struggle to see how such a case could be prosecuted without eyewitness testimony?

Maybe it shouldn't be then. Just because a crime happened doesn't mean someone has to be punished for it irrespective of their actual guilt.

Do you want the punisher (tm)? This is how we get the punisher(tm).

> Just because a crime happened doesn't mean someone has to be punished for it irrespective of their actual guilt.

The failure to prosecute the case of someone attacking someone else in their home merely creates an incentive for people to enact more violence in response, because the system fails to ensure baseline safety and won't be able to hold them accountable for their revenge.

Punishment for violence against others is required to have a peaceable society.

You can argue what you are, but you're advocating for a world of anarchy and uncontested violence.

I can hardly see why respecting due process leads to anarchism or violence.

In your scenario, why wouldn't people take revenge on each other by making false accusations and getting their enemies thrown in jail?

It's not like putting people in prison is a peaceful thing. It's a very violent ordeal, it's just violence that the state perpetrates on your behalf. We have to be sure that it's worth while.

> I ended up trying a case where the accused had allegedly gone to the victim's home, knocked on the door, then barged in and punched the victim a few times when he answered the door. Hospital records and photos meant no-one disputed the victim had been punched. But what evidence was there that the accused was the attacker? Only the eyewitness testimony of the accused and his wife.

Because we're not talking about an abstract situation of sans evidence accusations, and because that's already a problem -- even without eyewitness testimony being admissible.

We're talking about the different rates of bad things we can have in real compromises we can enact for real situations: if you go so far towards never convicting the innocent that you never convict violent home invasions, society will break down.

You can think that's justice, I was just connecting the choice with the outcome.

As counterpoint, one could say that you're advocating for a world in which it is better to sentence an innocent person for a crime than to let the crime go unpunished.

That's quite different from the sort of eye witness testimony discussed in the article, though. The odds of accidentally incorrectly identifying your own spouse is pretty damn low.

Your case more about he-said-she-said than 'our own eyes lying to us'. Completely different issues.

You misread the OP, the husband was attacked by another person, the husband and the wife identified the attacker as the defendant.

He didn't, the OP says 'accused and his wife', which really reads as though he admitted it.

I'm sure what was meant was 'victim and his wife' though.

Correct.

>Unless perhaps people put CCTV cameras all around in their own homes, and to me that doesn't sound like an improvement in civil liberties.

The typical solution that doesn't infringe on personal freedom is to let people posses whatever arms and armaments they want in their own homes (letting people do whatever in their own homes is good policy in almost all contexts). If there's even a 1:10 chance of getting swiss cheesed this sort of "inflict violence on a random stranger in their own home" type of crime falls of a cliff. It's also a preventative solution because it makes the difficulty of successfully committing the crime harder compared to a reactive solution (like CCTV) that only makes the chance of getting caught afterward higher.

As an aside I'm expecting a record low number of home invasions in Texas for 2019 considering the two high profile cases where the people breaking in got a little bit more resistance than they were expecting.

Eye witness evidence can be very deceptive but it can be very accurate as well. Many a time witnesses have provided key information to investigators. But, I agree, it cannot (should not) be used as primary evidence in criminal cases. Hippocampus or not, it’s not consistently reliable.

"Plane crash investigators mostly ignore eye witness testimony using it only at best to guide the investigation"

I once read that it is common for aircraft accident eyewitnesses to erroneously claim the airplane was on fire before it struck the ground.

>Also, this article doesn’t mention the cross or other race effect[2] which is yet another compounding factor.

The other race effect was explained in the fifth paragraph of the article. It's mentioned repeatedly.

As you see, we really cannot trust the testimony of our own eyes. We've known this for over a century. But we still do it.

Rashomon would have been shorter and simpler if they'd had surveillance cameras, though not an excellent movie anymore.

There's some debate about the accuracy of "flashbulb memories".

One side thinks that such traumatic events create durable, accurate memories. The other thinks that such events merely create a strong feeling of confidence that the memories are durable and accurate.

https://journals.sagepub.com/doi/abs/10.1111/1467-9280.02453

I've experienced this; my impressions of what other drivers were doing is rarely borne by the dashcam video.

I am reminded of a particular "Olde Time Radio" episode of 'The Shadow' which involves a trial; a person of [political? judicial?] importance is accused of receiving or at least being approached with blackmail//bribery. Fatal evidence was reel-to-reel film video with accompanying reel-to-reel audio played in tandem ["high-tech" at the time, remember].

Lamont Cranston & Margo Lane are in the courtroom audience for plot reasons and Lamont has issue with how the voice audio [doesn't quite] sync up with the lips on the video despite potential technology sync issues.

"Believe only half of what you see," Lamont paraphrases//declares to Margo, "and nothing you hear."

I'll let folks enjoy the episode to the end and I realize that the quoted paraphrase is likely a paraphrased quote in and of itself, but many decades after both I've still remembered the phrasing.

-/-/-

I've encountered this myself personally when trying to be a personal witness to someone escaping bounty hunters through a hotel bathroom window. In trying to be A Good Citizen, literally being voluntarily in the face of a cigar-chewing dude with sunglasses, and intellectually knowing what "tricks" might be played on me [e.g. "tall, blond hair?" -- "no, short and curly black!"] I still made rookie mistakes in basic reality-telling. And this was live-and-in-person shortly after the event. "Eye witnesses are unreliable" as sometimes reported, indeed.

"I was certain, but I was wrong"(2000) https://www.nytimes.com/2000/06/18/opinion/i-was-certain-but...

I think there is also a racial component to it. We focus on features that enable us to distinguish people in our social circle. For example, for some groups of people hair color and eye color are not really distinguishing (a majority of the group have black hair and brown eyes). It would be interesting to see how many of the inaccurate identifications were cross-racial.

Sex too.

The Innocence Project has overturned hundreds of wrongful convictions, around 99% men.

I have been talking to some chinese researchers working on facial recognition who claim that 3d infrared video is already more accurate than eyewitness identification, for chinese people, in china.

I thought about this problem recently when I was listening to To Kill a Mockingbird. Quoting Atticus:

> “...but in the absence of eyewitnesses there’s always a doubt, sometimes only the shadow of a doubt.”

One word: corroboration.

Reply on news.ycombinator.com