The National Registry of Exonerations

[Dcdrac]

http://www.law.umich.edu/special/exoneration/Pages/about.aspx

Just came across this and it proved an interesting read.

[Draca]

For all exonerations, the most common causal factors that contributed to the
underlying false convictions are perjury or false accusation (51%), mistaken eyewitness identification (43%) and official misconduct (42%) – followed by false or misleading forensic evidence (24%) and false confession (16%). The frequencies of these causal factors vary greatly from one type of crime to another.

http://www.law.umich.edu/special/exoneration/Documents/exonerations_us_1989_2012_summary.pdf

http://www.law.umich.edu/special/exoneration/Documents/exonerations_us_1989_2012_full_report.pdf


I expected eyewitness identification to be high, but it does take me by surprise that pergury is a factor in 51% of the cases and official misconduct in 42%.

[theprestige]

http://www.law.umich.edu/special/exoneration/Documents/exonerations_us_1989_2012_summary.pdf

http://www.law.umich.edu/special/exoneration/Documents/exonerations_us_1989_2012_full_report.pdf


I expected eyewitness identification to be high, but it does take me by surprise that pergury is a factor in 51% of the cases and official misconduct in 42%.

That doesn't surprise me at all, actually.

What did you expect to be a major cause of false convictions, if not false accusations?

As for official misconduct, I wonder if that's just egregious and willful misconduct, or if it also includes any and every possible form of official mistake and misbehavior.

[Rolfe]

I read up some interesting stuff on eyewitness identifications some months ago, and there was surprising information.

If there are people in the room with the witness who know which picture in a photospread is that of the suspect, the chance that the witness will pick that picture goes up dramatically.

If witnesses are given books of photos to leaf through, and one picture appears two or three times interspersed with others, the witness becomes likely to pick that one because it now looks familiar.

Witnesses who have been asked to produce a photofit or artist's impression are subsequently less likely to pick the right person out of a line-up, because their perception becomes skewed towards the way the photofit/artist's impression turned out, which inevitably will be different from the real person.

Witness ability to recognise someone seen once declines dramatically with time, becoming no better than chance after 11 months. If the original circumstances of seeing the suspect were mundane and unthreatening, it's worse than that.

Even under the best conditions, with the witness getting a good look at the suspect under abnormal conditions (that is, realising they should be trying to remember this person) and only a short delay before testing, accuracy is only about 70% to 80%.

There are individual examples of witnesses becoming absolutely convinced that a certain person committed a crime they witnessed, and yet subsequent evidence proved that was impossible. The most widely-cited one was a woman who became convinced that the police suspect was her rapist, to the point she was absolutely adamant about it. DNA evidence later exonerated that man and implicated someone else. The woman accepted the DNA evidence was correct, but said she couldn't cope with it emotionally - she still saw the original suspect whenever she thought about the rape. The police questioning had caused the original suspect's face to replace that of the real rapist in her memory of the event.

The adversarial justice system is partly to blame. Having an eye-witness declare "it was him!" is hellish effective in court. Getting an expert in the psychology of facial recognition to point out the problems with believing that just sounds like desperation on the part of the defence. It's a real problem.

Rolfe.

[Dcdrac]

wasn't there an experiment with a man in a gorilla suit in a video and some students asked if they saw him or not?

[Rolfe]

Yes, I remember seeing that. It's probably on YouTube. Viewers were asked to concentrate on counting something in the clip (to do with people catching a basketball), and most people concentrated on that to the point where they didn't notice this freaking great gorilla walk through the game.

I think the problem of facial recognition is a bit different though. I know I'm terrible at it, and would always refuse to commit myself if accuracy was important. I'm always failing to recognise people I've seen before, or picking out the wrong person shortly after meeting someone for the first time. I have trouble figuring out what level normal people perform at, to be honest. But I'm damn sure some of the feats of recognition some convictions have been built on are just not credible.

Rolfe.

ETA, happy birthday Dcdrac, by the way!

[gerdbonk]

wasn't there an experiment with a man in a gorilla suit in a video and some students asked if they saw him or not?

The Invisible Gorilla (http://www.theinvisiblegorilla.com/gorilla_experiment.html)

[Draca]

Would you be a good eyewitness?

Selective Attention Test

http://injustice-anywhere.org/EyewitnessSkillsTest.html

[Beerina]

Mod rename link -- defeats purpose if you know about it beforehand!!!!!!!

The Invisible Gorilla (http://www.theinvisiblegorilla.com/gorilla_experiment.html)

If you need a rationale, imagine someone revealed a magic trick in the magician section...

Oh and rename the link here, too, of course.

[Rolfe]

I've had a few experiences myself that worry me as regards eyewitness identification.

Some months ago I drove up to the back entrance at work on a Saturday morning, when only myself and one technician are scheduled to be on duty. I glanced at the window, and was certain I saw a trainee technician there, who isn't yet on the Saturday rota. I immediately assumed he'd come in to shadow the duty technician to be trained to participate in the rota.

I went into the building, and the only living breathing person there was the duty technician. Who is a red-haired Scotsman. The trainee I thought I saw is a dark-haired Italian. They look nothing like each other. I cannot imagine I saw Sean through the window and thought it was Stefano. It seems more likely the entire thing was simply a trick of my visual cortex.

But suppose I hadn't gone in, and for some reason it had been important whether Stefano had really been there that morning or not. I could have given sworn testimony he was, testimony which would have been sincere and likely to be believed in the absence of evidence to the contrary.

Then a couple of days ago, I was visiting a garden attraction. I saw a woman coming into the refreshment area, and was certain I recognised her but was unsure from where. She had dark grey hair done in an unusual French knot style, and a slightly receding chin. She sat down next to me with no sign of recognition, and I continued to rack my brains. The hairstyle was exact. The age and general shape of face were extremely close. I thought perhaps this woman was slightly heavier than the person in my memory, but I was unsure.

It was only after she'd got up and gone away I placed the memory. A woman I have attended three choral workshops with, but haven't conversed with much. I still don't know if it was the same person. It would have been very easy to be convinced it was the same woman though.

Later the same afternoon a man came up to me and said, hey, aren't you the vet from the VI lab? Quite right. I didn't recognise him at all. He said, "we just got back into sheep" and I immediately knew which farm he belonged to. But I couldn't have picked him out of an ID parade for a pension, even though he's twice brought dead animals to me for examination, and it takes a good ten minutes to take a full history for these submissions.

But people say "I never forget a face!" and stick to their guns. How can they be so sure?

Rolfe.

[Rolfe]

Mod rename link -- defeats purpose if you know about it beforehand!!!!!!!

If you need a rationale, imagine someone revealed a magic trick in the magician section...

Oh and rename the link here, too, of course.


It's no use Beerina, the text on the page linked to explains it up front anyway.

Rolfe.

[crimresearch]

I read up some interesting stuff on eyewitness identifications some months ago, and there was surprising information.

If there are people in the room with the witness who know which picture in a photospread is that of the suspect, the chance that the witness will pick that picture goes up dramatically.

<SNIP>

The adversarial justice system is partly to blame. Having an eye-witness declare "it was him!" is hellish effective in court. Getting an expert in the psychology of facial recognition to point out the problems with believing that just sounds like desperation on the part of the defence. It's a real problem.

Rolfe.
The mutation of the adversarial justice system into a win at any costs spectator sport, is greatly to blame.

[casebro]

Darn. I was expecting a Monty Python sketch.

And I didn't even get the Spanish Inquisition.

I guess I'll go back to my lumber jacking, if that's OK.

[I Ratant]

Yes, I remember seeing that. It's probably on YouTube. Viewers were asked to concentrate on counting something in the clip (to do with people catching a basketball), and most people concentrated on that to the point where they didn't notice this freaking great gorilla walk through the game.

I think the problem of facial recognition is a bit different though. I know I'm terrible at it, and would always refuse to commit myself if accuracy was important. I'm always failing to recognise people I've seen before, or picking out the wrong person shortly after meeting someone for the first time. I have trouble figuring out what level normal people perform at, to be honest. But I'm damn sure some of the feats of recognition some convictions have been built on are just not credible.

Rolfe.

ETA, happy birthday Dcdrac, by the way!
.
There was a gorilla in that clip???? I missed it the first time also.:p
And can almost never remember a face for any length of time.
The number of bogus convictions based on official "misconduct" is downright appalling.
Nancy Grace has had a few of her convictions as an AG in Georgia overturned for exactly that.
The history of such is awful in this country, with political aspirations by the prosecutor driving the need for a conviction.. any conviction, no matter who is harmed.

[I Ratant]

Darn. I was expecting a Monty Python sketch.

And I didn't even get the Spanish Inquisition.

I guess I'll go back to my lumber jacking, if that's OK.
.
Just stay out of the damn bars!

[Draca]

That doesn't surprise me at all, actually.

What did you expect to be a major cause of false convictions, if not false accusations?

As for official misconduct, I wonder if that's just egregious and willful misconduct, or if it also includes any and every possible form of official mistake and misbehavior.


I've read the Innocence Project's contributing causes of wrongful conviction list before:

•Eyewitness Misidentification
•Unvalidated or Improper Forensic Science
•False Confessions / Admissions
•Government Misconduct
•Informants or Snitches
•Bad Lawyering

http://www.innocenceproject.org/understand/

I was expecting Eyewitness testimony to come out as the number one contributing factor. I was aware that false confessions cause a lot of them. The IP doesn't list Misconduct until number four and doesn't list Perjury / False accusation at all. Although maybe they have some of that under Informants and Snitches.

How awful that someone would lie about something that would put someone in jail for life or execute them. I can think of several cases where that is the case unfortunately.

Amanda Knox had several false witnesses. One park bench witness was used by the authorities in more than one trial as a witness.

Ryan Ferguson had TWO lying witnesses. They were the only evidence he was convicted with and BOTH have now recanted their testimonies.
http://www.groundreport.com/US/Ryan-Ferguson-seeks-new-trial-to-prove-innocence/2945518
http://www.change.org/petitions/please-grant-ryan-ferguson-a-new-trial-or-freedom

[Draca]

I read up some interesting stuff on eyewitness identifications some months ago, and there was surprising information.

If there are people in the room with the witness who know which picture in a photospread is that of the suspect, the chance that the witness will pick that picture goes up dramatically.

If witnesses are given books of photos to leaf through, and one picture appears two or three times interspersed with others, the witness becomes likely to pick that one because it now looks familiar.

Witnesses who have been asked to produce a photofit or artist's impression are subsequently less likely to pick the right person out of a line-up, because their perception becomes skewed towards the way the photofit/artist's impression turned out, which inevitably will be different from the real person.

Witness ability to recognise someone seen once declines dramatically with time, becoming no better than chance after 11 months. If the original circumstances of seeing the suspect were mundane and unthreatening, it's worse than that.

Even under the best conditions, with the witness getting a good look at the suspect under abnormal conditions (that is, realising they should be trying to remember this person) and only a short delay before testing, accuracy is only about 70% to 80%.

There are individual examples of witnesses becoming absolutely convinced that a certain person committed a crime they witnessed, and yet subsequent evidence proved that was impossible. The most widely-cited one was a woman who became convinced that the police suspect was her rapist, to the point she was absolutely adamant about it. DNA evidence later exonerated that man and implicated someone else. The woman accepted the DNA evidence was correct, but said she couldn't cope with it emotionally - she still saw the original suspect whenever she thought about the rape. The police questioning had caused the original suspect's face to replace that of the real rapist in her memory of the event.

The adversarial justice system is partly to blame. Having an eye-witness declare "it was him!" is hellish effective in court. Getting an expert in the psychology of facial recognition to point out the problems with believing that just sounds like desperation on the part of the defence. It's a real problem.

Rolfe.


Absolutely Rolfe. The mind is scarily malleable. There is a case that cleary shows what you describe in a real life case. Jennifer Thomspon was raped and identified Ronald Cotton. He was later cleared by DNA. They now tour giving speeches on the subject together. It's also an amazing story of forgiveness.

Here is their story on 60 Minutes - part 1-2
http://www.cbsnews.com/video/watch/?id=5153451n
http://www.cbsnews.com/video/watch/?id=5153459n&tag=contentBody;storyMediaBox

Eyewitness Identification - Getting it Right
http://www.injustice-anywhere.org/EyewitnessTestimony.html

[casebro]

One thing to keep in mind is that those rates apply to exonerations, not convictions in general.

So what is the overall exoneration rate?

[geni]

http://www.law.umich.edu/special/exoneration/Documents/exonerations_us_1989_2012_summary.pdf

http://www.law.umich.edu/special/exoneration/Documents/exonerations_us_1989_2012_full_report.pdf


I expected eyewitness identification to be high, but it does take me by surprise that pergury is a factor in 51% of the cases and official misconduct in 42%.

I suspect that official misconduct is so high simply because its the easiest to prove. The requirement for law enforement to document everything means there is plently of written evidence to dig through.

[Professor Yaffle]

Eyewitness evidence (extremely weak) was a major part of the case that put Sam Hallam away.

http://www.hackneygazette.co.uk/news/sam_hallam_s_distress_to_discover_photos_of_his_la te_father_proved_he_was_not_at_the_crime_scene_1_1 386565

http://www.channel4.com/news/sam-hallam-and-miscarriages-of-justice

[Cuddles]

How awful that someone would lie about something that would put someone in jail for life or execute them.

While that is awful, as Rolfe has noted it's not usually actual lying. People just aren't as good at recognising or remembering things as they think they are. The trouble is that no matter how much you explain that and no matter how much evidence you have to support it, people will almost never believe it applies to them. "Sure, people are generally terrible at remembering but I never forget a face".

Just read some of the threads here about UFOs and other claims to see how insistent people are that they remember incredibly minor things correctly even after being shown overwhelming evidence that they're almost certainly wrong. Put them in a situation where they have the opportunity to make sure a murderer is locked up and they're hardly going to be more open to accepting it. They know they saw what they saw, and no amount of science is going to convince them otherwise.

[Rolfe]

Absolutely Rolfe. The mind is scarily malleable. There is a case that cleary shows what you describe in a real life case. Jennifer Thomspon was raped and identified Ronald Cotton. He was later cleared by DNA. They now tour giving speeches on the subject together. It's also an amazing story of forgiveness.

Here is their story on 60 Minutes - part 1-2
http://www.cbsnews.com/video/watch/?id=5153451n
http://www.cbsnews.com/video/watch/?id=5153459n&tag=contentBody;storyMediaBox

Eyewitness Identification - Getting it Right
http://www.injustice-anywhere.org/EyewitnessTestimony.html


I think that's the case I was thinking of. I read a long article about it some time ago. The problem there is that the witness is really certain and really sincere, and neither the authorities or the jury or they themselves realise how likely it is that they are sincerely mistaken.

Since you've mentioned the Knox/Sollecito thing, maybe I can mention the identification evidence that was crucial in Megrahi's conviction. In that case the witness (Tony Gauci) was anything but certain. Bizarrely, the judges ruled that this meant he was more likely to be right, because clearly he wasn't just blindly asserting what the police wanted. The possibility that he was uncertain because it wasn't the same guy but the police wanted him to say it was, didn't seem to occur to them.

There's an article about the eyewitness identification in that case here (http://www.vetpath.co.uk/lockerbie/photoid.pdf). Clothes found to have been packed in the suitcase with the bomb were traced to a small shop in Malta, and amazingly enough the shopkeeper remembered selling them to a man who spoke Arabic, nine months previously. His description was quite detailed, but it was mostly about size and body shape rather than face - as you'd expect from someone who sells clothes for a living. He made a photofit and and an artist's impression, and said they were both OK likenesses, but the weird thing is they look like two different people.

The cops then started showing him photographs of faces, trying to get a better handle on what the man looked like, and Tony picked out several pictures as "resembling" the purchaser. More than two years after the transaction, they were able to get him to pick out a very bad likeness of Megrahi from a photospread, using a version of the "Clever Hans" trick. The fact that Megrahi turned out to be nothing like the original description as regards height, age, build, body shape and even skin colour didn't seem to matter.

Over ten years after the transaction they finally got Tony Gauci and Megrahi together for a live identity parade. Megrahi looked nothing like the original description, or the photofit or the artist's impression, or the bad passport photo Tony had picked out eight years previously, but Tony still hesitantly picked him out of the line-up. Of course, everyone and his dog knew what the "Lockerbie bomber" looked like by then, so that wasn't really hard. But still he said, it's not actually the man but it looks like him.

The first problem is that the cops seemed to be treating it like a game they had to win, and getting Tony to point to the right photo was the "win". They never stopped to think, overall, is it actually likely this really was the man? Then of course the prosecution grab it and run with it. An eyewitness identified Megrahi as the man who bought the clothes in the bomb suitcase! Case closed!

Tony, by the time of the trial, had been convinced by the cops that they were certain the man charged was the bomber, and it was up to him to make sure the "bad man" didn't get away with it by doing his part. You can almost hear the conflict when you read his statements. He had no clue whether or not Megrahi bought these clothes - he had completely forgotten what the purchaser looked like by 1999, and probably by 1991 actually. (Except his 1989 description of the six-foot-tall, burly, dark-skinned 50-year-old was on file.) But he genuinely wanted to help, and he and his brother knew there was a reward of millions waiting for them if he got it right, too.

But none of this seemed to be considered by the judges. The prosecution produced an eyewitness identification, and that was fine.

There are three expert witness reports linked in the article linked to above, and all of them start with a general exposition of why we should be very very wary of eyewitness identification of strangers, particularly strangers of a different race. They make sobering reading.

Rolfe.

[wald1900]

I've read the Innocence Project's contributing causes of wrongful conviction list before:

•Eyewitness Misidentification
•Unvalidated or Improper Forensic Science
•False Confessions / Admissions
•Government Misconduct
•Informants or Snitches
•Bad Lawyering



I was expecting Eyewitness testimony to come out as the number one contributing factor. I was aware that false confessions cause a lot of them. The IP doesn't list Misconduct until number four and doesn't list Perjury / False accusation at all. Although maybe they have some of that under Informants and Snitches.



Y’know what surprised me most about this summary is the prevalence of police or prosecutorial misconduct; “Official Misconduct” played a role in 42% of the registered cases. Wow! I was pretty tuned into false confessions, eyewitness misidentification and even bad forensics, but actual misconduct? Sure, I knew there had to be some, but not anywhere near this order of magnitude.

This got me thinking about the sheer number of people still in prison as a result of crooked cops and crooked prosecutors. While it would be nice to gather all of these monsters in one prison and throw away the key, such measures would do nothing to help people who continue to serve time as a result of past misconduct. In the interest of getting these victims out, what would you guys think of a proposal for limited amnesty to police and/or prosecutors who self-confessed to misconduct?

It could work something like this…..

For some defined period lasting, say, 90 days, any police officer, prosecutor or outside contractor (e.g. employees of independent forensics labs) who has engaged in “misconduct” can self-report without fear of criminal or civil charges. After the 90 days has lapsed, they become subject to a whole host of punitive laws aimed at sending these idiots to prison with mandatory sentences at least equal to the time their victims have spent in jail and subjecting them to unlimited personal civil liability. In other words, “come forward now or reap the whirlwind”. As a condition of the amnesty, the offending official would have to testify to their wrongdoing at any retrial of the criminal victim and tender a public apology. The good news would be that they don’t have to go to prison or lose their homes and retirement savings. The bad news would be that don’t get to be cops or DAs (or attorneys?) anymore.

On the victim side of the equation, the fact that misconduct has occurred does not guarantee that the defendant is innocent. However, a self-confession by an official involved in the victim’s case would automatically guarantee the victim to a speedy retrial. Assuming most of these would result in an acquittal the law could also set generous standards for compensation such that, although under the amnesty the victims would lose their right to civil relief against the offending official, they are made whole (or whole-enough) by the state.

To the extent that a given cop or DA has, over the course of their career, participated in the arrest and prosecution of literally hundreds of other defendants, the majority of whom (presumably) are actually guilty, it can be assumed by statute that if the bad guys self-confess that they’ve come clean on all sins and so, avoid retrials on every case they’ve ever been involved with over the course of their careers.

In cases where a cop or cops have worked with a DA to fudge a conviction, it would create a fascinating dynamic among the co-conspirators…a classic prisoner’s dilemma. If just one of a group of co-conspirators self-confesses while the others choose not to, those who pass on the amnesty risk severe criminal and civil penalties under the new, harsher laws (which, by the way would have no statue of limitations). In cases of conspiracy, the conditions of the amnesty would almost guarantee that at least one (and probably, all) of the co-conspirators would come forward to avoid prosecution. Can you imagine the cockroaches that would come running out of the woodwork in Perugia if such a deal was put on the table for Perugia PD and members of the prosecutor’s office? I betcha that on the 91st day we’d all be sitting on our sofas eating popcorn and watching Amanda’s long-lost interrogation video.

Whadaya think?

[Rolfe]

Just read some of the threads here about UFOs and other claims to see how insistent people are that they remember incredibly minor things correctly even after being shown overwhelming evidence that they're almost certainly wrong.


That reminds me of a very public example of this that happened just after Jean Charles de Menezes was shot by police on the London underground. There were a lot of eyewitnesses, and several of them came on TV and radio shortly afterwards to relate what they said they saw. There were various stories of the dead man running from police, someone said there were wires hanging from his collar, almost everyone said he was wearing a heavy jacket.

One witness in particular gave a lurid account of de Menezes allegedly running along the platform, glancing behind him at the pursuing police, looking like "a hunted fox". Then he tripped and fell on the platform, and the police threw themselves on top of him and he was shot in the head five times.

None of that happened. de Menezes, wearing a white T-shirt with only a short denim jacket, unbuttoned, over it - and no wires, not even an iPod, had only run a short distance when he saw the train already in the station. He boarded and took his seat in a leisurely manner. He sat down and began to read the free paper he had picked up as he walked into the station. The police entered the train, at which point the cop who had been tailing him (who was already seated on the train) stood up and indicated which passenger was the suspect. The armed team advanced on de Menezes, who stood up as he realised he was the object of their attention. One policeman pinned his arms to his sides and forced him back into his seat, at which point one of the other policemen shot him in the head at least seven times.

Was the witness simply fantasising? Who knows. But if it hadn't been for the sheer number of eyewitnesses, and some CCTV cameras, that wildly inaccurate account could have become accepted fact.

Rolfe.