TODAY   |  April 21, 2014

Rossen Reports: How reliable are eyewitnesses?

Do you trust your memory? In this segment of Rossen Reports, national investigative correspondent Jeff Rossen puts eyewitness memory to the test, setting up a social experiment where people witness a crime and are asked to identify the perpetrator, with some surprising results.

Share This:

This content comes from Closed Captioning that was broadcast along with this program.

>>> we're back now at:42. this morning on "rossen reports", wii kicking off a two-part series called "can you trust." if you witnessed a crime, could you accurately pick out a suspect? today, national investigative correspondent jeff rossen is putting our memory to the test. i will say i will fail this right now.

>> see you later then, take care. we all like to think we're astute, except for matt, always paying attention. let me test you at home right now. close your eyes . the person sitting right next to you, what color are they wearing? what's the shirt look like? do you remember? okay, open your eyes . if you got it right, great. but i'm guessing so many of you got it wrong, even when the stakes are high, like after a crime, it's important. police rely on eyewitnesss to pick out the suspect, and when they're wrong, innocent people are thrown in jail. so can you trust your memory? the morning, our hidden cameras are rolling. you're watching a crime in progress. this man just stole a woman's purse, and he's making a run for it.

>> did you guys see my bag? somebody took it?

>> all these people saw the thief.

>> he did.

>> he walked away.

>> but would they be able to pick his face out of a lineup?

>> he's the one.

>> this fellow right there.

>> you actually chose the wrong guy.

>> wow.

>> research shows when it comes to faces, our memory can fail us. and in real crimes, eyewitnesss often make mistakes. jennifer disart is an expert in photo lineups.

>> for people who choose and say yes, i recognize that person, but 40% of the time they're incorrect.

>> that's a startling number.

>> extremely.

>> so we set up a social experiment inside this bar and grill in the new york suburbs. wierg it up with hidden cameras and sending in two producers. jove vanna is our victim and josh is the thief. okay go to the bathroom now. jo vanna steps away, leaving her handbag on the bar.

>> here comes our suspect.

>> as josh goes for the bag, he knocks over some silverware, purposefully drawing attention on his way ouhis way out. we ran the same scenario in front of dozens of customers.

>> he walked out the door, i swear to god.

>> he took it?

>> maybe your height, dirty blondish.

>> time to test their memory.

>> high, guys. we're from nbc, the "today" show. this is actually a social experiment.

>> we asked the witnesses to pick the perp out of a photo lineup. for you at home, can you pick out our thief? that's him. right there.

>> well, he basically had -- it was like a rough beard.

>> he had facial hair.

>> nope. he's clean shaven .

>> i feel it's this person here.

>> wrong again. one after the other, they seem to pick everyone. except our perp. this witness is sure he can spot him.

>> if i see him again, i definitely am going to be able to recognize him.

>> do you recognize him here?

>> i think it will be here.

>> wrong.

>> just going through the visual in my head, like photographic memory .

>> a lot of people believe that memory might work like a video recorder. close their eyes and replay things. but that's really not how it works.

>> and there's more. experts say if witnesses don't see the suspect in the lineup, they'll still pick the next best option. like a multiple choice quiz. i'm going to show you some photos.

>> okay.

>> in fact, watch what happens when we take our perp's photo out entirely.

>> he's too dark. he's too dark. i don't know. something about him. he's too light. so he's the one.

>> nope. he's innocent. we figured this next witness would be better.

>> he dropped some utensils, i turned around, i saw him.

>> he's a retired police detective .

>> this is him.

>> how sure are you?

>> about 100%.

>> confident. but still wrong.

>> in the end, every single witness who made an i.d. in both lineups picked the wrong guy.

>> the consequences are severe because once the eyewitness makes a mistake, it's very, very difficult to stop the process. and a conviction is likely to happen.

>> police around the country are trying to reduce these eyewitness mistakes, looking at new procedures for lineups. for example, they're going to start showing one photo at a time instead of a multiple choice , like is employed now. and making sure the officer doesn't ask leading questions to get the result they want. but there is some good news here. we have some tips on our website right now,, to make you a better witness.

>> who are you?

>> by the way, let me turn around for a minute. what color tie am i wearing?

>> brown and blue.

>> you just copied her. you already said you have a bad memory.

>> yeah, but i was watching because i knew i'd be quizzed.

>> that's a great witness there.