In examining convictions later overturned by DNA evidence, Dr. Gary Wells discovered that 90% of those convictions were based on eyewitness evidence. In one case, 5 separate eyewitnesses wrongly identified the same person.
Eyewitness testimony is most often called into question due to visibility conditions. However, there are many other ways in which sensory perception is fallible, ranging from the amount of attention during the initial observation to the innate inability of most people to identify facial features accurately. The way in which lineups and other police and courtroom procedures are conducted can also lead to an inaccurate identification.
Presence of a weapon
When a crime is committed with a weapon, eyewitness accuracy falls. This is called the weapon focus effect.” Dr. Elizabeth Loftus has used eye-tracking technology to demonstrate that the presence of a weapon draws the visual focus of the eyewitness away from everything but the weapon. This includes the perpetrator’s face.
People are not good at identifying faces. They are worse when trying to identify a cross-racial face, regardless of which race they belong to. Neither racial animosity nor cross-racial familiarity make any difference.
When the eyewitness is presented with lineups in which the criminal is not present, the eyewitness is likely to misidentify one of the members of the lineup as the perpetrator. In some cases, merely being presented with a police suspect can be enough to cause eyewitness misidentification.
It is believed that, in the absence of proper instruction, an eyewitness chooses a person out of the lineup who most resembles his memory of the perpetrator, rather than a person who exactly resembles the perpetrator. Wells found that even a simple instruction to the witness that the perpetrator may not be present in the lineup is enough to dramatically reduce the number of false identifications.
To reduce the number of misidentifications resulting in false convictions, the Department of Justice has published a set of best practices for conducting police lineups. Another recommendation is to use known innocent fillers who match the victim’s description of the subject, including any unusual hairstyles or other distinguishing features. Under no circumstances should more than one suspect be included in the same lineup.
A Minnesota pilot study has found that presenting photos or individuals sequentially rather than together increases the accuracy of eyewitness identification. Presenting the photos of possible suspects sequentially reduces eyewitness pressure to choose a suspect out of the presented lineup. It also lets the eyewitness consider each photo individually and comparing it to memory, rather than comparing the members of the lineup to each other for similarity to the perpetrator.
Preparation for cross-examination
The more time passes, the less reliable is eyewitness memory. Studies have found that accuracy falls sharply within 20 minutes of the original event, and keeps on doing so until the 2nd day. At the same time, eyewitness memory is vulnerable to contamination through inappropriate feedback. However, where an eyewitness is braced to testify in court, his certainty in his memory increases.
Dr. Amy Bradfield has found that an eyewitness who is prepared for cross-examination does not have a more reliable memory of the perpetrator, but thinks he does. Even learning that another witness has identified the suspect increases the certainty of other eyewitnesses. The increase in certainty is actually higher in cases where it later turns out the eyewitness was wrong.
Jurors often treat eyewitness testimony as compelling evidence which outweighs any other form of evidence. However, it is highly unreliable, especially in cross-racial cases which involve a weapon. A long time interval between the crime and trial also increases the amount of memory loss and contamination.