Everywhere from NextDoor to Quora to YouTube  comments boxes, we hear people say: “Eyewitness testimony is the most unreliable evidence.” And they credit “scientists” for proving it.

That’s BS.

Sorry to discolor the decorum of Mind Matters News, but that’s my reaction.

Triggering my outburst is the recent article from ScienceInsights headlined: “Why Science Tells Us” Not to Rely On Eyewitness Accounts.” Once again, a formless abstraction, “science,” is dictating unquestionable truth. Let’s see about that.

What are the problems eyewitness testimony is said to create?

It gets criticized from several directions. First, human memory isn’t a video camera. It doesn’t record; it reconstructs. We rebuild events from what we have noticed, stored, and later recalled. Errors can occur at every stage. A witness may not notice important details and memories can fade or change over time. Later recall can be influenced by added information.

Image Credit: Photographee.eu - Adobe Stock

Stress makes these problems worse. Faced with a frightening crime or other dangerous situation, a person’s attention may narrow. If a weapon, raging fire or other startling factor is present, people may focus on that factor instead of on other people or details.

Memories can also be altered after the event. News reports, social media, conversations with other witnesses, being asked leading questions, or picking up subtle reactions from investigators can introduce new information. Of course, alcohol, drugs, age, and limits on attention or working memory may further reduce accuracy.

One point that is especially important in legal cases and even casual situations is this: The fact that a witness strongly claims that a memory is correct does not mean it is accurate. Feedback from investigators and other people, repeated questioning, repeated exposure to a suspect or person involved, and the passage of time can make witnesses increasingly certain of what they believe they saw or heard. Such witnesses are not dishonest, it’s just that their memories have become inaccurate.

These memory problems occur. Look carefully, however, at how subjective they are. For example, post-event factors can alter memories — but how much? And what is lost? Memories fade over time, sure, but it matters whether we fail to recall the key facts vs. other details.

Memory flaws can affect life-and-death matters

Great concerns rightfully arose when people were convicted of murder or other serious crimes based upon eyewitness testimony, and then years later DNA evidence showed they were innocent of those crimes. Reported statistics estimate that eyewitnesses incorrectly identify suspects around 25% of the time when looking at photos or people in a police line-up. Upwards of 75% of wrongful convictions overturned by DNA evidence during the periods studied involved eyewitnesses falsely identifying the suspect.

Image Credit: Photographee.eu - Adobe Stock

It became paramount to the criminal justice system to figure out whether eyewitness testimony is reliable enough to support convictions for serious crimes. Researchers found that people in both laboratory and real-life experiments often would not recall a significant number of details. The worst situations occurred when, using memory alone, eyewitnesses incorrectly identified as the possible criminal a person “they saw” and ended up supporting his or her conviction.

The National Academy of Sciences weighed in after reviewing academic sources. It concluded in 2014 that human vision and memory impose fundamental limits on eyewitness accuracy. Memories can be forgotten, reconstructed, updated, and distorted without the witness realizing that this has even happened.

Regrettably, a combination of headlines and soundbites swept over popular culture. People became convinced that eyewitness testimony is unreliable, period.

Car crashes vs. sock drawers

Survivors and witnesses to car crashes saw, heard, maybe felt, shocking, extraordinary events. Investigators of car accidents will ask them many questions, digging down into details that the people would not necessarily have cared about when the tires screeched, metal smashed, glass shattered, and bodies bled. When asked, people’s recall can be incomplete and possibly inaccurate about some lesser facts, but they’ll all agree there was an accident which featured the key elements they saw. The same evidence problems affect investigations of murders, armed robberies, rapes and kidnapping.

Now consider this real conversation at home:

Me: Honey, where are my patriotic Christmas stockings?

She: In the dresser, fourth drawer down on the left.

Me: How do you know that?

She: I saw them when we were looking for sweatpants.

Me: That was two years ago!

She: … And your point is?

Nothing wrong with my wife’s memory! In your own life, you and people around you remember a zillion things just fine. (Senior citizens: different things, but still zillions.) My wife’s an eyewitness and she’s right practically every time. Daily life supports eyewitness memory as the best and sometimes only source of some information.

Eyewitness testimony can be the best

Critics devalue eyewitness testimony mostly when it appears in criminal prosecutions where the strongest evidence against the defendant is the single witness saying she “saw him do it.”  (For millennia, single witness accusations have been disfavored.)  Instead of discarding the eyewitness account as naturally untrustworthy, however, researchers have found ways to help make sure that only highly reliable eyewitness information is used.

Image Credit: PheelingsMedia - Adobe Stock

We start by understanding memory as an information-bearing system whose reliability varies with identifiable conditions. Sometimes the witness has an excellent opportunity to observe, and then provides a correct immediate account, makes a prompt and highly confident identification under fair conditions, and supplies details independently confirmed by other evidence. Instead of a knee-jerk response that “eyewitness memory is too fallible,” current research supports a better principle:

Preserve the first memory. Protect it from contamination from outside sources, interviewer bias, and targeted-loaded and leading questions. Test it fairly. Immediately record the witness’s confidence in the accuracy of the memory. Then evaluate the strength of the evidence. Don’t assume eyewitness testimony is either inherently trustworthy or inherently unreliable.

Prof. John T. Wixted and colleagues’ paper, “Rethinking the Reliability of Eyewitness Memory” (2018), offers more technical details about obtaining trustworthy eyewitness information (quoted and paraphrased here):

Eyewitness testimony is likely to be very reliable if the following conditions are met:

(a) Witnesses were not previously exposed to distorting or contaminating information,

(b) the witness’ memory is being probed for the first time,

(c) witnesses are not “tricked” into providing desired information (e.g., by biased lineups or suggestive interview questions),

(d) the witness doesn’t claim accuracy for uncertain memories, or explicitly reports his or her level of confidence, and

(e) the investigator is sensitive to the witness’s level of confidence, by relying on high confidence responses much more than low-confidence responses.

When such conditions are met, eyewitness memory is likely to be reliable — both in the laboratory and in the field — in tests of either recognition or recall. When such conditions are not met, eyewitness memory may be unreliable but that is hardly because of a faulty memory system.

Eyewitness Gold

Concerns about eyewitness identification of a perpetrator shouldn’t make the whole debate focus on photos and lineups. Eyewitness evidence is often much broader. In an armed robbery case, for example, eyewitnesses can supply vital facts for further investigation and evidence collection. They might accurately see and hear the following:

• two people entered the store

• one carried a rifle

• one stayed at the entrance

• the other approached the counter

• an argument with the checkout clerk occurred

• three shots were fired

• the men left in a blue truck

Eyewitness testimony can provide what no laboratory technique can completely replace: The most important evidence about an event may come only from the human beings who were actually there.

Recalling an entire event isn’t an all-or-nothing proposition. Being accurate or mistaken about one feature does not mean that every other remembered feature is right or wrong. Each fact deserves individual consideration. Being exactly correct in identifying another human is only one way a witness is valuable.

Next time someone belittles eyewitness testimony as “the least reliable,” we have the work of Prof. Wixted and colleagues who show otherwise. Honestly, without eyewitness information, what is left? Surveillance cameras, “lie detectors” (sic), and AI predictions?

The trend against eyewitness reliability seems part of the ideology that undermines finding objective truth.