Eyewitness Testimony and the Failings of Memory

Eyewitness Testimony and the Failings of Memory December 9, 2013

Sharon Hill has an article in the Huffington Post about the many failings and misunderstandings of how our memory operates and the implications for the use of eyewitness testimony in criminal trials. She starts with a fact that is now very well-established in science:

If there is one fact that I wish we could all accept early in life, I would vote for drumming in the idea that memory is not like a tape recorder. If we learn this truth about the human mind, we could avoid so much trouble.

Memory is constructed. Pause a moment and let that sink in.

Memory is not objective, it is constructed by our own brains. It is not burned, or ingrained, or seared into it, as much as we would like to think that is the case. The truth is less precise, uncertain, and disturbing.

It’s also very susceptible to changing over time and being influenced by those who seek to probe our memories, like therapists and police interrogators. This has become the textbook example:

Some of the most egregious examples of memory failure (and worse, fabrication of false memories) was during the time in America’s history known as the Satanic Panic period of the 1980s. Many people were accused of heinous crimes relating to Satanic ritual abuse in homes, day care facilities, and schools. Scores of innocent people were convicted on testimony from witnesses who felt certain they were telling the truth about being tortured or seeing torture of others. These convicted people went to jail for decades. Their lives were destroyed. Their families were decimated.

It is too difficult for me to imagine sitting in jail, convicted of a terrible crime that just didn’t happen and not being able to do anything about it. Witnesses lied — they didn’t even know they lied. They believed those false memories really happened due to manipulative psychotherapy procedures that created a story out of whole cloth. It is too easy to warp a memory into something different, wrong or fantastic, but this is done repeatedly, daily, and can cause great harm.

She links to an article about Elizabeth Loftus, whose work on the fallibility of memory is starting to have an effect on how criminal cases are handled in some places. We need much more of this.

"“I used to watch it like a rocket ship when I put out a beauty.”I'm ..."

Trump: Saying Something Bad About Something ..."
"It occurs to me that the last time a President spoke this way (while in ..."

Trump Aims a Vile, Racist Tweet ..."
"To continue off my past post: There is a simple answer to the guessing game ..."

Partly Agreeing with Clarence Thomas on ..."
"Like, what's the alternative? Start to play the guessing game "ok, did they mean for ..."

Partly Agreeing with Clarence Thomas on ..."

Browse Our Archives

Follow Us!

What Are Your Thoughts?leave a comment
  • Memory can also determined by our expectations. Apparently it is very common for witnesses to airline crashes reporting the aircraft is on fire before the crash, even if it isn’t.

  • From what I’ve read, there is a growing consensus that memory works by assembling “archetypes.” You have only a very few images of, say, the taste of bacon: when you think about this or that time you ate bacon, a kind of “memory assembly daemon” in your brain uses those few images over and over again. False memories are easy to create, as all you have to do is convince the memory assembler to use different archetypes.

  • Wylann

    And sometimes it’s not faulty memory, but outright lying. (I’m thinking of my recent court case involving a rather careless driver who simply couldn’t tell the truth.)

  • eric

    She links to an article about Elizabeth Loftus, whose work on the fallibility of memory is starting to have an effect on how criminal cases are handled

    …which is extremely depressing, considering Loftus has been publishing studies on the fallibility of memory for literally decades – since the 1970s.

  • David C Brayton

    I’ve seen some videotapes of interviews of people that have later recanted their testimony in the Satanic cases. It is amazing how suggestive the interviewers were. The interviewers almost certainly wanted to elicit certain responses. It was as if they were training puppies–lots of positive reinforcement for certain answers.

    I was dumbstruck that these ‘witnesses’ were ever allowed to testify. And was outraged that the juries believed their testimony after seeing these ‘interviews’.

    I realize that hysteria can take hold of an entire community–think the Salem witch trials–but in the science and technology age, it shouldn’t happen.

  • cptdoom

    I studied with Saul Kassin, who has worked closely with Dr. Loftus, as an undergraduate, first in a course entitled “Psychology and the Law,” which looked at the realities of how human minds work and how the judicial system operates (e.g., many states do not allow jurors to take notes during trials, yet note-taking leads to better outcomes, particularly because trials tend to happen in crimes with indeterminate and confusing evidence). I was shocked to learn not only how faulty eyewitness testimony is, but also that such testimony is the most compelling to juries, compounding the damage. And there are so many ways in which eyewitnesses may be wrong – the weapon effect, in which attention tends to be drawn to a weapon, not the criminal; the cross-racial identification problem, in which we are all less able to distinguish features among people who are different from us racially and ethnically; and, of course, the susceptibility of eyewitnesses to deliberate and inadvertent suggestions from law enforcement, which not only changes people’s memory, but can make them more confident in those erroneous memories.

    I also learned how to be a better eyewitness – concentrating on clothing and characteristics such as height and build, which are harder to confuse. When I was subsequently held up at gunpoint, the police woman who took my statement was impressed with the details I had gotten, but also a bit shocked that I stated flatly I would not be able to identify the guy in a line up. In this case, however, the police cut a break because the guy robbed about 10 people in a three-week period, and every time hid his gun in a bright yellow plastic bag from a local video store. Every one of us identified the bag and when the guy was arrested, the link among the crimes was relatively easy to make.

  • tbp1

    I heard Elizabeth Loftus speak some years ago and was very impressed. At the time she was traveling with bodyguards because she had been threatened by people unhappy with her work on recovered memory in abuse cases. There was a kind of cottage industry at the time (is there still?—I really don’t know) of people claiming to be able to dig out repressed memory of childhood abuse, and there was money to be made by therapists and lawyers. It was very rarely, if ever, true but people with an ax to grind (or money to make) didn’t want that word to get out. I don’t know if she ever actually suffered an attack, but there was a credible enough threat for her to have the guards.

  • voidhawk

    The thing which drove this home to me was during James Randi’s exposition of a medium, in which a medium gave a reading to a subject who was asked to asess the accuracy of the reading. The subject said somewhere between 80-90% and said that the medium had only said 2 or three names before getting it right. When Randoi actually read out how many names had passed the Medium’s lips it was closer to 30 or 40 names. The subject simply did not remember them.

  • colnago80

    Re tbp1 @ #7

    The McMartin case in Tinseltown is the prime example of the fraudulence of the recovered memory hucksters. Aside from the waste of taxpayers money, the lives of members of the McMartin family were ruined for, despite their acquittal.


  • wscott

    My earliest childhood memory is of the Dumbo Ride at Disneyland when I was 2. And it’s a completely manufactured memory. I had been told about that trip and how much I loved the Dumbo ride since forever, and I had built a false memory around it. In my 30s, I came across a picture of the Dumbo ride in a Disney book I’d owned as a kid, and only then did I realize my meory was based on that photo, not the ride itself. (In hindsight, the fact that the Dumbo ride in my memory didn’t move should’ve been a clue…)


    So yeah, memory is (sometimes) bullshit.

    There was a kind of cottage industry at the time (is there still?—I really don’t know) of people claiming to be able to dig out repressed memory of childhood abuse, and there was money to be made by therapists and lawyers. It was very rarely, if ever, true but people with an ax to grind (or money to make) didn’t want that word to get out.

    Based on the recovered memory crusaders I’ve encountered, I think most of them truly believe they’re doing good work. I’m sure there are a few hucksters out there, but my impression is the majority believe they are genuinely helping people. Their resistance to the truth about recovered memories strikes me as more justification biases than outright deceit. Of course the results are the same, but no need to ascribe malice where basic human irrationality will suffice.

  • ursamajor

    Back in the early 70s I was reading the essays in a college psychology text and the articles on memory were very shocking. The studies (I have forgotten the details of who and when – probably done in the late 50s) involved having people review sketches of scenes then later recall as many details as possible. The first thing that struck me was how little people could recall after even short waits. The other was the patterns to how people reconstructed what they saw. One image was of a town square with a pharmacy in the middle of a block. Universally the study participants (probably all university students) moved the pharmacy to a street corner. Another picture was of the interior of a bus and among other things going on a white man had a knife and was threatening a black man. Again, in memory the knife was usually in the hand of the black man.

    There went any trust I had in eyewitness reports.

  • jaytheostrich

    It boggles my mind that witness testimony is still given any credibility in courts these days. I tend to assume that about one third of my memories are wrong or false, with the amount of dreams I have, books I’ve read, etc. It’s a little like that fact that the brain never stores the source of it’s memories, which is why I never read articles from the Onion or similar sites. I don’t need false information like that cramming itself in there too!

  • Oh, joy. More ammo for the abuse apologists and rape apologists, backed by science.

  • prodegtion

    And this is why testimony is completely inadmissible in the court of science.

  • prodegtion


    Yeah, because who needs actual evidence for accusations of rape, right?

  • davidbrown

    And smart criminals know how to take advantage of this. When the Stopwatch Gang robbed banks, they would put a large, red rubber ball on the end of the aerial of the getaway car they were driving. A few blocks away they would get rid of the ball. Witnesses at the bank couldn’t remember the make of car, its colour, or how many people were in it, but they all agreed it had a big, red rubber ball on the end of the aerial.

  • You know what’s disheartening? I’ve read some work suggesting that even when mock jurors are well-prepped with knowledge of the problems with eyewitness testimony, there is no appreciable impact on their judgment as compared to those without prep. Now it may be that because it’s a mock jury, everyone knows that all the testimony is fictional, but I fear that the bias is so strong that even in real trials the bias still drives belief in the accuracy of sincerely presented testimony. They sound and seem truthful, so what they’re saying must be true. Something like that.


  • CSB

    @13 @15:

    I can see how this might be a problem if, for example, a rapist chooses as their defense that that the sex was consensual. Still, I find it very difficult to accept that the increase in false negatives w/r/t guilt could be enough to even consider preserving the unwarranted deference that eyewitness testimony is given.

    The sad fact is, those who wish to discredit individuals who claim to be victims of rape or domestic assault would have a wealth of tactics to choose from regardless. They could claim the victim to be lying (either directly or by omission), for example, making a specific example of testimony untrustworthy independent of how much trust can be placed in eyewitness testimony as a category.

    Or, to put it more simply, this doesn’t give apologists any tools they haven’t already used many times before.

  • I’ve had a lifelong flashbulb memory of a man who looked like Mr. Greenjeans standing over our bathroom sink, shaking like a leaf. There was blood all over him. This was in the bathroom of a house we moved out of just before my 4th birthday.

    I recently asked my father if he knew what that was about, and he immediately recalled. There was a terrible car wreck in front of our home. A woman passenger was killed and my father brought the husband into our bathroom to clean up and try to calm him. My mother reports the same.

    Now, surely my parents relayed that story to others when I was very young, but I only have that flashbulb image. So is that an impression of what I actually saw, reconstructed again and again over the years, or did I build the image hearing my parents discuss the accident with others? Could be both? I’ll never know, but I can see it clear as day.

  • tbp1

    @13: How does pointing out the unreliability of eyewitness testimony constitute rape or abuse apologetics? It just means that we have to be careful about evidence.

    Here’s why I’m skeptical about eyewitness testimony. I have excellent memory about many things, but I am, personally, really really bad about facial recognition. I am perfectly capable of failing to recognize someone I’ve met literally dozens of times, even my own students if I see them out of context (like at a grocery store or at the movies). At the beginning of each semester I actually apologize in advance should I fail to recognize anyone outside of class. I know not everyone is as bad as I am, but I always wonder how many are but just won’t/can’t admit it.

    And a very concrete example: some friends of ours that we have lost touch with over the years had a cute story about one of their kids. I think that my wife and I just heard them relate the story. My wife insists that we were actually present when the cute incident occurred. We are both convinced we are right. I suppose we could try to track down our old friends and see how they remember it, but it’s not actually terribly important, and who knows if their memory would be any more reliable than ours so many years later. At any event, it makes me very doubtful people’s memory. Not necessarily their honestly, just their memory.

    Give me good circumstantial evidence (fingerprints, etc.) any day over eyewitness testimony.

  • DrVanNostrand


    It’s probably not as bad as it seems. My understanding is that most rapes are by people known to the accuser. The science about how bad eyewitness testimony is tends to be about people identifying strangers they’ve never seen before. If I saw a random dude mug someone, I’d have a hard time identifying them in a lineup, even in the best of circumstances. If I saw my friend Greg mug someone, I could identify him really well (provided it was well-lit and I was reasonably close, etc…) In any case, bad eyewitness testimony is often the major cause of conviction for exonerated death row inmates and plays a role in all kinds of unjust convictions. We simply can’t ignore that fact or the science behind the phenomenon.

  • barry21

    Bryan Fischer/David Barton type Christians love to mention that the Biblical standard for the death penalty is corroborating testimony from two eyewitnesses. They hold up the standard as evidence of Biblical wisdom and fairness.


  • nancymannikko

    One of the saddest cases I’ve personally seen of constructed memories was when I served as a juror for a homicide trial. The pretrial questioning two children had been subjected to left both of them (an 11-year-old and a 7-year-old) convinced they had been eye witnesses to their mother being shot. The defense attorney was able to pick enough holes in the narrative to make it clear to the jury that it was physically impossible for the kids to have been present in the room where the shootings occurred, but those poor kids are probably going to spend the rest of their lives convinced they saw something horrific. Granted, waking up to the sound of gunfire and coming out of the bedroom to find your mother dead is pretty horrible in itself, but it’s still a notch better than being sure you saw the actual shooting.

  • Two words. “Tookie” Amirault. Railroaded into an 18 year incarceration on the testimony of children younger than 5 yo; testimony that was blatantly coerced by “therapists” and Martha Coakley’s prosecution team.

    If there was justice in MA, Martha Coakley, the idiot that handed Scott Brown a Senate seat would be serving the same term as the one she engineered for Amirault, his sister and his mother. I wouldn’t piss on Coakley if she was on fire.

  • khms

    In the 1970s, I read (among other stuff) Perry Mason novels. Unreliability of eye (and ear, and …) witnesses is a point made in a lot of those. Circumstantial evidence is much more reliable (but usually needs interpretation, which of course is also fallible).

    So I learned about that back then.

  • eric


    And this is why testimony is completely inadmissible in the court of science.

    We in science have the luxury of benig able to spend a lot more time and money coming to a belief decision, when the data is poor we can choose not to come to a decision with pretty much no penalty, and we have the further luxury that (in general; there are exceptions) an incorrect decision will not have an immediate, devastating impact on some person’s livelihood. These methodological luxuries are why we can afford very stringent standards of evidence. Courts must make decisions that send people to jail. They have to make them in a relatively short period of time and with relatively limited resources*, and they don’t really have the choice of ‘not coming to a decision’ if the data is poor.

    And despite all those luxuries, it looks like – at least in some fields – we in science have serious problems with irreproducibility. There are probably a lot of innocent people in jail, but I don’t think anyone believes that 75% of the people in jail were wrongly convicted – yet recent findings in science show that only 25% of (studied) papers could be reproduced. IOW we are arguably doing way, way worse than the court system in terms of ‘getting the right answer,’ despite our superior standards of evidence. The reason this is not news is simply because most published scientific results are comparatively inconsequential. Collectively science is incredibly important, but your average ‘bit’ of published results is not. In fact so much of published results are inconsequential that it took us years to figure out how much of it was wrong.

    Now, I’m a big science supporter. I think that for empirical discovery it’s a spectacularly good methodology, inarguably the best we’ve got. I hope that there is a trend in courts and juries to weigh scientific evidence more and more and weigh eyewitness testimony less and less. But with all that benig said, the court system’s goals are different. Their methodological limitations are different. Their standards of evidence are different for some pretty good reasons.

    *Before you cite some anecodotal long and expensive court case, go search for the average time taken for a criminal case in your county. I bet you will find that its on the order of 1-5 days.

  • Georgia Sam

    Combine this with revelations about police interrogation methods that tend to produce false confessions (see for example the latest issue of The New Yorker), and it paints a pretty bleak picture of a wrongly accused person’s chances of getting justice.

  • gshelley

    Some of the comments are a little depressing, pretty much along the lines of “no, my memory (or that of the witness in whatever case they are interested) is reliable” and all the evidence is wrong

  • colnago80

    Re democommie @ #24

    The same Martha Coakley who claimed that she couldn’t prosecute Cardinal Law because being an accessory after the fact was not a crime in Massachusetts. She let that piece of filth scurry away to his current bolt hole in the Vatican.

  • freehand

    WMDKitty, if a woman says she was raped by her boss, then barring substantial evidence casting doubt on her testimony, I would believe her. But if she identified a stranger in a police lineup, I think that a man whom police suspected the most* has been fingered, and he may or may not be the actual culprit. If I or a member of my family were hurt, I would not want the real criminal to escape. I am concerned not with believing her but rather in correctly apprehending her real attacker.

    Years ago I saw a trespasser at the lab I worked in, and we discovered missing tools soon after. The police took me downtown and ran me through various mugshots, and really wanted me to finger one person in particular. I only saw him with the sun behind him, however. I told them the could be the man I saw, but there must be many more that would fit the observations (medium height, slender black man with short hair, face unseen). It was a poor metropolitan area with a predominately black population. There were probably 100,000 men within ten miles who might have been the one. I found the whole procedure disturbing. They searched his apartment anyway, and found the tools. But they wanted a conviction more than justice.

    * Or simply wanted him to be convicted the most. Or he was at hand.

  • Crip Dyke, Right Reverend Feminist FuckToy of Death & Her Handmaiden

    The truth is less precise, uncertain, and disturbing.


    Fucking grammar, how does it work?