Eyewitness Identification – How Reliable Is It?

I’ve recently been engaged in an “inter-editorial discussion” concerning the accuracy of eyewitness identification.  This stemmed from a comment I made on a blog post citing Innocence Project data stating that 75% of the IP DNA exoneration cases have involved incorrect eyewitness identification.  My comment was that the IP has data showing that eyewitness identification is wrong 75% of the time.  Well …. that may the case for this particular set of data (289 DNA exonerations), but it cannot be validly extended to eyewitness identifications in general.

So, how reliable is eyewitness identification?  I think the only thing we can say for sure is that we don’t know for sure, but we do know it’s not very good.  Three different studies from 1987 to 1998 (Wells, Huff, Cutler & Penrod) have determined that eyewitness identification is wrong anywhere from 35% to 60% of the time, and one study even determined that it was wrong in 90% of cases studied.  That’s a huge range of results (and even 35% is not good), and this is because the accuracy of eyewitness identification depends on SO MANY things:  lighting, distance, amount of activity at the scene, the presence of a weapon, the fear of personal harm, the visual acuity of the observer, time delay from observance to identification, the methods used for conducting police lineups, cross-racial effects, age and gender of the observer, and on and on.  Added to this is the fact that human memory has been shown to be “malleable” – it changes over time in response to a wide range of influences, and people can be subject to the “power of suggestion”.

The US justice system gives great credence to eyewitness identification, and an eyewitness identification will even trump an airtight alibi in court.  Given this, and the fact that eyewitness identification reliability is not good – we have a problem.  Now the good news is that the problem is beginning to be recognized, and ‘some’ steps are being taken to help rectify this – like sequential, double-blind police lineups.  The bad news is that the problem still exists, and probably always will.  But more good news is that research into the issue continues, some of which is cited below.

If you would like a real “eye opener” on the subject, I recommend the book Picking Cotton by Ronald Cotton and Jennifer Thompson-Cannino.  The book details an instance in which the victim had close, lengthy, one-on-one contact with the perpetrator, and still got the eyewitness identification wrong.

And here is some more reference material for you:

Here is the Innocence Project report on lineups and eyewitness identification:


Here is a link to an article by Dr. Marc Green, who is a human factors expert, and has studied eyewitness misidentification.


Prof. Gary Wells of Iowa State University has been studying eyewitness identification for years, and is a recognized authority.  Here is a link to his website:


The Eyewitness Identification Reform Litigation Network, which is comprised of a number of post-conviction innocence organizations, including the Innocence Project and the National Association of Criminal Defense Lawyers, maintains a website devoted to the issue of eyewitness identification.  Here is a link:


6 responses to “Eyewitness Identification – How Reliable Is It?

  1. Thanks, Phil, for this comprehensive look at research on eyewitness accuracy. An important study of CODIS DNA database matching records indicated a consistent 25 percent error rate in victim identification of the perpetrator in sexual assault cases. Even if eyewitness error is “only” 25 percent that makes eyewitness identification very unreliable and certainly not enough standing alone to convict beyond a reasonable doubt.

  2. Pingback: Defending People » Blawg Review #317

  3. so, nancy. What if you have corresponding eyewitness accounts leading to let’s say, a rape. 2 women saw a man walk into the room, and the third was assaulted. The man has an alibi, which isn’t taken as “Unreliable”, the testimony is. So we let him walk? Eyewitness testimony is essential to convictions, though shouldn’t be the only means of conviction.

  4. Stephen – I just offer up the example of Roger Dean Gillispie, recently exonerated here in Ohio, of rape by the way. The only thing that convicted him was the eyewitness identifications of the two victims who were abducted at the same time and together. But guess what? It’s been proven he didn’t do it. You can read about “Dean’s” case in Nancy and Jim Petro’s book False Justice.

  5. It’s really a cool and helpful piece of information. I’m happy that you just shared this useful info
    with us. Please keep us informed like this. Thanks for sharing.

  6. Pingback: Eyewitnesses Are Often Wrong | Wrongful Convictions Blog

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s