Based on the book Unfair: The New Science of Criminal Injustice by Adam Benforado.
For more detailed information, read the book.
The Fallibility of Eye Witnesses & Memory
Eye witness testimony is perhaps the most persuasive evidence in a trial. Yet, it can be unreliable and easily biased. Witness certainty has little relationship to accuracy. In a lineup that includes the actual perpetrator, witnesses fail to pick out anyone about one-third of the time and pick out the wrong person a third of the time. In a lineup without the perpetrator included, people select someone about half the time. Of the first 250 DNA exonerations in the United States, 190 of them appear to have involved mistaken identifications.
We rarely remember information that we do not consider relevant, and many factors in a crime do not seem relevant until after the incident. We remember the gist of an incident but not the details. We remember events that elicit strong emotions, but not the surrounding facts. And our memories are changed by subsequent experiences. We fill in fragmented memories with highly specific details to form a narrative influenced by our expectations and what we want to believe.
The accuracy of eye-witness identification is strongly affected by distance, lighting of the event, age of the observer, and differences of age and race between the witness and the suspect. High levels of stress, anxiety and physical exertion also decrease accuracy of memory. We tend to focus our attention on what we see as the threat (e.g., the gun) and not on the identity of the perpetrator.
Composite drawings of perpetrators are only slightly better than random choices in a lineup, and constructing the image can alter a witness’s memory. Seeing a composite can affect later identifications by witnesses. People assume a photo array contains the perpetrator and will pick someone unless told the perpetrator may not be included. Also, accuracy of a photo array drops precipitously after about seven days.
Eyewitness testimony and lineups can be effective investigation procedures, but they must be done carefully because they can easily be tainted. Comments by other people, especially poorly trained police officers, can influence witness memory making it less accurate but more confident. For example, a live lineup is the most accurate type of lineup IF all the people in the lineup match the initial perpetrator description, and IF the witness has not been shown a person in the lineup previously or subjected to comments about their choice. In-court eyewitness identification is the most powerful testimony to jurors in spite of being highly unreliable. Scientific research shows that eyewitnesses misidentify innocent people 30 percent of the time. We would not accept that level of error anywhere else. We must correct our bias toward the accuracy of eyewitness testimony.
Science is already being used in many jurisdictions to improve witness accuracy. That includes recording identification procedures, having someone other than the primary investigator handle the process, using sequential lineups that are recorded, requiring the prosecution to prove reliability if challenged by the defense, using cognitive interviewing techniques (establishing rapport, using open-ended questions with few interruptions, and avoiding leading questions) in interrogations, and educating jurors on the actual reliability of witnesses. Witnesses are very valuable, even essential, in many cases, but jurors must understand their limitations and fallibility.