Mistaken eyewitnesses are the most common and often the most compelling piece of evidence in wrongful convictions.

In the book "Actual Innocence" by Barry Scheck, Peter Neufeld, and Jim Dwyer (pg 318) is stated, "In 2000, the Innocence Project reconstructed 74 cases in the U.S. of the 79 exonerations in North America to determine what factors had been prevalent in the wrongful convictions."

• Mistaken eyewitnesses were a factor in 82%
• Police misconduct 50%
• Prosecutorial Misconduct 45%
• Tainted or fraudulent science 33%
• Defense lawyers falling down on the job 32%
• False confessions 22%
• Snitches or informants 19%

Another study found that out of 205 wrongful convictions, the main cause in 52.3% of the cases was eyewitness misidentification ("Convicted But Innocent: Wrongful Conviction and Public Policy"; C. Ronald Huff, Arye Rattner, Edward Sagarin). The next most common cause was perjury by witnesses at 11%.

Tom O'Connor, Assistant Professor of Justice Studies at North Carolina Wesleyan College explained the error rate in his article by stating that "several studies have tried to estimate the error rate, but it appears to jump all over the place. An early study by Borchard (1932) reported that eyewitness error contributed to 45% of the wrongful convictions in his sample. Buckhout (1974) regarded it as the most common cause of wrongful conviction. Rattner (1983) reported that more than 52% of the wrongful convictions he studied involved eyewitness error. For capital cases only, Bedau and Radelet (1987) found a 16% error rate. In the process of making an identification from photographic lineups, Loftus (1979) found that 29% were erroneously chosen. Huff, Rattner & Sagarin's (1996) more recent survey of judges and attorneys found that 60% of judges and 84% of attorneys ranked "accidental eyewitness misidentification" as the most frequent cause of wrongful conviction. In 1999, the APA's Div. 41 (American Psychology-Law Society) reviewed a number of cases that were later overturned, and found an eyewitness error rate of an astonishing 90%."

Mistaken identifications range from ones in which the convicted bears a close resemblance to the perpetrator to cases where it is hard to understand how the mistake could have been made. For example in one case the victim described her attacker as being clean shaven. Later, the vitcim became convinced that her attacker wore a beard.

A large number of eye witnesses are simply cooperating with suggestive authorities and lineups. Prejudice and cross-racial identifications also increase the likelihood of misidentifications. Tom O'Connor claims that "other errors are the result of psychological trauma or shock experienced by the eyewitness, and these may be exacerbated by hypnosis sessions intended to help the eyewitness remember or calm down. And, in numerous other ways eyewitnesses mislead themselves; it is obvious that our criminal justice system has an imperfect way of handling eyewitness identifications and/or balancing the rights of the accused with the inherent power of eyewitness testimony."