Chapter 17: Evaluating Eyewitness Testimony of Adults

Welcome to Last Minute Lecture.

This free chapter overview is designed to help students review and understand key concepts.

These summaries supplement not replaced the original textbook and may not be redistributed or resold.

For complete coverage, always consult the official text.

Usually when we think about evidence in a criminal trial, we picture something solid, you know?

Yeah.

Like DNA, a fingerprint, maybe a fiber from a jacket, things you can just put in a plastic bag and hold up to the light.

Physical evidence just feels, I don't know, absolute.

Right, yeah.

It doesn't change its mind.

It doesn't get confused by poor lighting or anything like that.

Exactly.

But then you look at a case like the one from Los Centro, Washington.

This was in January of 1993.

A 36 -year -old housekeeper is working alone down this really secluded driveway,

and two men just barge in.

Which is an incredibly terrifying scenario to begin with.

Oh, absolutely.

One has dark hair, one is blonde.

They assault her.

They actually tie her to a dining table with an electric cord, and this is key, they put electrical tape over her eyes.

There's this horrific, chaotic trauma.

And when it's all over, the physical evidence they collect doesn't immediately point to anyone at all.

Wow.

So they have almost nothing physical to go on.

Right.

The entire case basically hinges on something incredibly fragile, her memory.

And you know, the foundation of the investigation begins to crack almost immediately, long before anyone even realizes a mistake is being made.

Yeah, because within days, she sits down with a sketch artist.

She admits she didn't really get a good look at the blonde man, but she tries her best to describe the dark -haired one.

But she's struggling, like, with the specific details of his nose and mouth.

Which is completely understandable given the trauma.

Of course.

But the sketch artist, instead of just leaving it vague,

simply adds, and I quote, A nose that fits with the face.

So that sketch goes public, tips start coming in saying it looks like a local logger named Alan Northrup, the police put together a photo montage, the victim looks at it, and nothing.

She doesn't identify him.

She doesn't identify him at all.

And I have to say, in a strictly scientific process, failing to identify the primary suspect in a photo array is a glaring red flag.

It's a strong indicator that you need to look elsewhere.

But they don't.

They keep digging into Northrup's circle instead.

Weeks pass, and they hear about a friend of his, Larry Davis, who sort of fits the description of the blonde man.

So they show the victim a new montage with Davis in it.

She points to Davis's photo and says, that's not him, but that's the one.

Wait, that's not him, but that's the one?

Later, she actually testifies she couldn't identify his face, but his neck looks familiar.

And the police log that as a positive identification.

I know.

Soon after, they run live lineups.

She identifies both Davis and Northrup.

In 1993, they are convicted and sentenced to over 20 years in prison.

And it wasn't until 2010, after they had spent 17 years behind bars, that DNA testing of 27 different pieces of crime scene evidence completely excluded both men.

They were fully exonerated.

It is just an absolute tragedy.

And if you are a college student diving into forensic psychology, this is the exact kind of case that keeps you up at night.

So for today's deep dive, you and I are opening up the Handbook of Forensic Psychology, fourth edition, chapter 17.

A really crucial chapter.

Definitely.

Our mission today is to walk you through the exact legal standards, the psychological theories, and the forensic practices used to evaluate adult eyewitness testimony.

We're going to look at the science of why human memory fails us and really how the justice system accidentally wagonizes those failures.

Okay, let's unpack this.

Let's do it.

To understand how the victim in our opening story could make such a tragic mistake, we really have to start with the blueprint of how humans make decisions when they are entirely sure of the answer.

Right.

And to understand that blueprint, we turn to something called signal detection theory, or SDT.

Interestingly, this didn't actually come from a psychology lab originally.

It came from World War II.

Oh, really?

Yeah.

Imagine a radar operator staring at a screen.

A blip appears.

They have to make a quick decision.

Is that an enemy bomber or is it just a flock of birds?

Their decision actually rests on two completely independent components.

Okay, what's the first one?

The first is the informational component.

This is just the raw data available.

How clear is the blip?

Is there static on the screen?

In our forensic context, this is the actual memory trace the witness has in their brain.

Did they get a good look?

Was the lighting good?

But the raw data is never enough on its own, though.

Exactly.

The second part of the equation is the decisional component.

This is the subjective threshold, like the internal bar you set for making a choice.

If the radar operator decides to wait until they are 100 % certain it's a bomber, that's a very strict conservative decisional threshold.

Got it.

And if they ring the alarm with the slightest weird shape on the screen, that's a very loose liberal threshold.

Tot -ton.

So, applying this to a police lineup, there are basically four ways this can play out, right?

First, the suspect in the lineup actually did the crime and the witness picks them.

That's a hit.

Second, the suspect is guilty but the witness says, nope, not them.

That's a miss.

And the other two outcomes happen when the police put an innocent person in the lineup.

If the suspect is innocent and the witness points to them anyway, that's a false alarm.

The worst case scenario.

Exactly.

Finally, if the suspect is innocent and the witness correctly says, they aren't here, that is a correct rejection.

Now, if we connect this to the bigger picture, you can see how manipulating that decisional component completely changes the outcomes.

How so?

Well, if you instruct a witness to be very cautious, shifting their decision criterion to be more conservative,

you successfully reduce the number of false alarms, meaning you protect innocent people.

But statistically, you also reduce the number of hits.

A few more guilty people might slip through because the witness was just too afraid to guess.

You know, it sounds just like taking a multiple choice exam where you kind of know the material, so that's your informational component, but you have to decide if guessing is actually worth the risk.

That's a great analogy.

Yeah, like if the professor says there's a huge point penalty for a wrong answer, your decisional threshold goes way up.

You leave it blank unless you are absolutely certain.

You might miss out on some points you could have gotten, but you avoid the penalty.

And a witness is constantly weighing those invisible penalties.

They want to help the police, they want the bad guy caught, they obviously don't want to look foolish, and they really don't want to ruin an innocent person's life.

All of that impacts where they set their internal bar for making an identification.

Right, so if the decision is built on those two pillars, let's look at what degrades the informational raw data before the police even arrive.

A lot of people, I think, just assume that law enforcement officers must be better at recognizing faces than the general public.

It's a common assumption, but the science definitively shows they are not.

Face recognition is basically a trait set in your teens and is incredibly difficult to train.

We are generally much better at recognizing someone we know well, but that introduces a really dangerous phenomenon called source confusion.

Which is when you recognize someone but can't place them.

Close.

You recognize a face, but your brain misattributes where you know them from.

The text cites this wild case in Seattle where two witnesses falsely identified a defendant as the perpetrator simply because they all used to work together briefly.

Oh wow.

Yeah.

The witnesses recognized his face, felt that really strong sense of familiarity, and just assumed, oh, he's familiar because he's the guy who robbed the store.

That is terrifying.

And this gets even more complicated when we cross racial or age lines, doesn't it?

It does.

Decades of research show we are much better at recognizing members of our own ethnic or racial group.

It's a phenomenon called own group bias.

And the numbers are pretty striking.

Witnesses are 1 .56 times more likely to falsely identify an out -group member.

That's a huge jump.

It is.

And this applies to age groups too.

Children under 10 and adults over 70 are prone to making significantly more false alarms in lineups.

The text mentions gender too, right?

Women are slightly better at face recognition generally.

But there was this fascinating study by Valentine at a horror labyrinth in London.

Oh, the haunted house study.

Visitors went through this terrifying, super stressful haunted house.

And the researchers found that female visitors were more likely to develop high levels of anxiety under that specific stress, which actually impaired their ability to identify a person they encountered in the labyrinth later on.

Right.

That study perfectly illustrates how extreme emotion disrupts encoding.

So what does this all mean?

Because honestly, I have to push back here.

My intuition says if someone is attacking me, if it is a literal life or death situation,

that intense stress would burn the attacker's face into my memory forever,

like a flash bulb going off in a dark room.

It's a very common misconception.

But cognitive psychology breaks stress down into two different modes.

The first is arousal mode.

This happens in less threatening situations.

Your physiological arousal elevates a bit, and your attention sharply focuses on the central details of an event.

Okay.

So in that mode, your memory actually gets better.

Exactly.

In arousal mode, your memory for the core event actually improves.

Well, that sounds like it supports the flash bulb idea.

It does, up to a point.

But a life or death attack pushes you past arousal mode and into what we call activation mode.

Activation mode.

Yeah.

This is high cognitive anxiety, true panic.

Your physiological arousal skyrockets to the point where you experience a catastrophic drop in memory performance.

Your brain is so overwhelmed by the biological imperative to just survive that it literally stops effectively encoding the details of the attacker's face.

That makes a lot of sense, actually.

Your brain is busy trying to keep you alive, not taking a photograph.

And if the attacker has a weapon, I imagine the brain locks onto that instead of the face.

Exactly.

It's called the weapon focus effect.

The weapon creates what researchers call attentional capture.

A gun or a knife is unexpected and obviously highly threatening, so your visual attention zeroes right in on it.

You might be able to describe the barrel of the gun perfectly, but it comes at the massive expense of the peripheral details, like the identity of the person holding it.

I'm guessing if we add intoxication into the mix, the whole encoding system just completely burns down.

Pretty much.

Alcohol predictably increases the likelihood of false identifications.

And marijuana is particularly interesting.

It significantly lowers the amount of detail a witness can initially recall right after a crime, though that difference in detail retrieval kind of fades after a week.

Interesting.

Yeah.

So all of these factors, lighting, distance, weapon focus, intoxication, extreme stress, they all dictate the quality of that informational memory trace left in the brain.

And once the crime ends, that memory trace stops forming.

From that second on, it is only degrading.

So the raw data, the memory itself is that fragile.

What happens when you introduce detectives into the mix?

Because this is where we see how police practices accidentally manipulate the witness's decisional threshold.

Right.

And standard police interviews are surprisingly inefficient.

They often rely heavily on close -ended questions like, was he tall or was the car blue?

Which sort of plants the idea of a blue car.

Exactly.

This puts the witness in a passive role and introduces suggestive concepts.

Psychologists vastly prefer the cognitive interview, which uses open -ended prompts to encourage natural memory retrieval without planting ideas.

Makes sense.

Also, if a witness forgets a detail and remembers it later, what we call reminiscent detail, juries often think they're lying.

But memory science shows those delayed recollections can actually be highly accurate because the brain just continues to process the retrieval cues over time.

Which really explains why composite sketches, like the one in our opening story, can be so destructive.

I mean, the artist literally just added a nose that fit.

Composites are notoriously problematic.

Laboratory performance of witnesses creating composites is usually just at chance level.

They rarely look like the actual perpetrator.

Wow.

And the act of making one actually contaminates the memory.

There's a famous study by Jenkins and Davies that showed if a stitch artist adds a mustache to a composite, the witness will often later remember the perpetrator having a mustache, even if he didn't.

The sketch overrides the original memory.

Elizabeth Loftus proved this too with her work on post -event information, didn't she?

She sure did.

She showed you can easily manipulate a witness's memory just by the way you phrase things after the fact.

You can make them remember a yield sign as a stop sign.

She even implanted a completely false childhood memory of taking a ride in a hot air balloon just by having family members suggest that it happened.

What's fascinating here is how this memory malleability collides with the police line up process.

A police line up is an incredibly intense social situation.

It inherently acts like a multiple choice test.

The witness walks in feeling this heavy social pressure, subtly absorbing the message that the right answer is somewhere in the room and the police are just waiting for them to pick him.

So the instructions given right before the line up have to be critical then for releasing that pressure.

Oh, absolutely.

The National Institute of Justice now requires police to explicitly instruct witnesses with the phrase, the perpetrator may or may not be present.

Just that one sentence.

Yes.

This seemingly small sentence gives the witness permission to make a correct rejection.

It significantly lowers the social pressure and adjusts their decisional threshold.

We also have to consider how the line up is physically shown, right?

The traditional way is the simultaneous line up, you know, six guys standing in a row.

But the text says the problem is, witnesses use a relative judgment strategy.

Right.

They look at all six faces and ask themselves, who looks most like the guy, rather than comparing each face individually to their internal memory.

So the solution is they show them sequentially instead, like one photo at a time.

That's the theory behind sequential lineups.

Yes, that it stops that relative comparison.

But recent data suggests it might just make witnesses more conservative across the board.

They make fewer false alarms, but they also catch fewer guilty people.

Wow, they see.

And what makes it worse is when police allow multiple laps, letting the witness view the sequence again after they've finished.

The science shows viewing two laps significantly increases errors because the faces become familiar just from the first lap.

So it's that source confusion again.

Exactly.

But the worst error of all is the repeated procedure, showing a witness a mugshot and then later putting that exact same guy in a live line up.

The witness picks him because they saw his mugshot last Tuesday, not because he's the actual attacker.

It artificially inflates the witness's confidence.

So police are artificially inflating a witness's confidence in the interrogation room.

What happens when that witness takes the stand and a jury looks them right in the eye?

Well, the legal system leans heavily on a 1977 U .S.

Supreme Court case, Manson v.

Braithwaite.

The court laid out five criteria for juries to evaluate if an eyewitness is reliable.

And one of the biggest criteria they emphasize is witness certainty, basically how confident the witness is on the stand.

Here's where it gets really interesting.

Because the science absolutely rejects that.

If psychologists know confidence can be manipulated by a single confirming comment from a detective,

I'm shocked the legal system still relies on the Manson criteria.

It's a classic example of the law relying on quote unquote common sense rather than empirical science.

Psychologists have tested the confidence accuracy relationship extensively.

The correlation is only about 0 .29.

0 .29.

Yeah.

In plain English, that means confidence only accounts for less than a 10 % overlap with actual accuracy.

A witness can be 100 % confident and 100 % wrong.

There has to be a better indicator than confidence, then.

There is.

Response latency.

How fast a decision is made.

Exactly.

Research consistently shows that immediate snap identification decisions are much more likely to be accurate than slow, agonizing, deliberate choices.

If a witness has to study a lineup for five minutes, the accuracy drops significantly.

I always cringe when a witness dramatically points to the defendant in the middle of the courtroom during a trial.

Like in the movies.

Oh, in court, identifications are considered by psychologists to be the most pernicious, dangerous form of evidence.

They lack all the protections of a fair lineup.

The defendant is literally sitting right there at the defense table.

It's incredibly suggestive.

Highly suggestive.

The prosecution relies on the illusion that the witness has an independent source of pristine memory from the crime itself.

But science tells us that pristine memory simply doesn't exist by the time a trial happens.

So when you, as a forensic psychology expert, are called to the stand, what are the boundaries of your testimony?

I mean, you aren't there to point at the witness and say, they are lying.

No, definitely not.

The expert must never give an ultimate opinion on guilt.

That invades the province of the jury.

My job is not to declare the witness accurate or inaccurate.

So what do you say?

My job is conditional.

I present the science and say, could the factors present in this specific investigation produce an identification of this person, even if they were entirely innocent?

Let's take all these scientific tools and legal frameworks we've just learned and apply them directly to the tragic case of Alan Morthrop and Larry Davis from the start of our deep dive, because tracing the compound errors reveals exactly how this happened.

The informational component was compromised early on.

Right.

The nose that fits in the composite sketch likely contaminated the victim's memory, giving her a false anchor right from the start.

And then when the victim failed to identify Morthrop in the first montage, police ignored that non -identification instead of treating it as evidence of his innocence.

Which led to the repeated procedures.

They showed the victim photo montages of the suspects and then later put those exact same men in live lineups.

And the only faces that were identical across the photos in the live lineups were Northrop and Davis.

So the familiarity she felt during the live lineup was almost certainly contaminated by the photos she'd seen weeks prior.

Her decisional threshold was just completely destroyed.

I mean, before the live lineups, she was tipped off by a friend at the jail and by the police themselves that suspects were in custody.

She walked into that room expecting the guilty party to be there.

Which is a huge problem.

The police actually told her they were doing the live lineup because she missed the suspect in the photo array.

That completely vaporizes the may or may not be present protection.

Absolutely.

And the text mentions a brilliant way to test how structurally biased those lineups were.

It was developed by Dub and Kirshenbaum in 1973.

It's called the mock witness evaluation.

How does that work?

If you want to know if a police lineup is biased, you take people who know absolutely nothing about the crime, you give them the victim's basic physical description of the them the lineup and ask them to guess who the police suspect is.

Oh, wow.

So if it's a fair lineup of six people, they should really only guess the suspect by random chance one in six.

So about 16 .6 % of the time.

Exactly.

And researchers actually ran this exact test on the photo montages used against Northrup and Davis.

And what happened?

The mock witnesses picked Alan Northrup 65 % of the time.

They pick Larry Davis 50 % of the time.

Are you kidding?

They guess the suspect 65 % of the time without even seeing a photo of the crime based purely on written description.

It completely exposes the structural bias.

This proved the lineups had what we call a low effective size.

The suspects stuck out like sore thumbs.

The fillers in the lineup were not providing any actual protection because they didn't match the original description well enough.

That is insane.

And this raises an important question for you, the listener.

If the science so clearly shows how these procedures manufacture false certainty, how many other convictions are currently resting on this exact same chain of compound psychological errors?

It's a truly chilling thought.

So to summarize our deep dive into Chapter 17 eyewitness memory is never just a playback of facts.

It is the product of two distinct forces,

the informational data encoded during the chaos of the event and the decisional thresholds that are so easily and often accidentally manipulated by police procedures.

It is a delicate, highly corruptible system.

I want to leave you with a final thought to mull over.

Your memory is not a video recording sitting on a hard drive in your brain.

It is a live reconstruction.

It is constantly being edited by your environment, your stress levels and the subtle suggestions of the people around you all without your permission.

That's very true.

So the next time you are absolutely 100 % certain about a memory from years ago, ask yourself who or what helped you edit it?

A powerful question.

We want to say a huge thank you to you, our college student listener, for joining us today.

We know you are prepping hard and the entire last minute lecture team sends a warm thank you for joining us on this special academic deep dive.

We'll catch you next time.