On an initial, uncontaminated memory test, eyewitnesses typically provide reliable evidence, and this is true even for the majority of wrongful convictions later reversed by DNA evidence. However many factors can affect this reliability including the stress of witnessing a crime, the presence of alcohol or drugs, the amount of time that has passed since the event occurred, and the degree to which the witness interacts with the defendant after the crime.
These issues can be difficult to assess without access to the witness, the defendant, and relevant documents. John Jordan, a professor of psychology at University of Michigan, has conducted studies on eyewitness testimony that have helped inform police practices regarding the collection of information from witnesses. His work has shown that simply taking notes during an interview can improve the quality of information provided by the witness later on in a process called "memory contamination." He has also studied how different types of questions affect the accuracy of eyewitness reports and has found that open-ended questions that allow for detailed answers tend to produce more accurate results than questions that only allow for "yes" or "no" responses.
Ultimately, the reliability of eyewitness testimony is based on how well investigators can account for possible biases when collecting information from witnesses.
Although witnesses are frequently certain that their recollection is correct when identifying a suspect, the malleability of human memory and visual perception makes eyewitness testimony one of the most untrustworthy pieces of evidence. A witness may misinterpret what he or she sees; also, other factors such as stress, alcohol, or drug use can affect a person's ability to remember events exactly as they happened.
In general, eyewitnesses are more likely to identify suspects who resemble the victim than those who do not. This is because people try to fit facts into existing knowledge structures or "mental models" of the world, and if an eyewitness sees someone he or she believes to be responsible for the crime, then it makes sense for him or her to say so. Factors such as gender, race, height, weight, age, and clothing style don't influence whether or not someone is identified by an eyewitness, but these characteristics can help determine whether the witness picks out the right person from a lineup or set of photos. For example, if there are several men with similar features, but only one of them is guilty, an eyewitness would be more likely to select him.
The accuracy of eyewitness testimony is highly dependent on the circumstances of the crime.
Even among the most confident witnesses, eyewitness testimony is a powerful kind of evidence for condemning the guilty because it is prone to unconscious memory distortions and biases. As a result, memory may be either amazingly accurate or surprisingly wrong. The two are indistinguishable in the absence of objective proof. In real life, eyewitness testimony is rarely exact; rather, it is usually a fair representation of what occurred.
In general, people are better at remembering events that affect them emotionally or physically than those that do not. This is called the "emotional significance" of events. People also have a tendency to focus on information that matches their beliefs or expectations. These are both examples of "biases" in thinking that can influence how someone recalls an event.
When eyewitnesses identify someone as the perpetrator of a crime, they are typically relying on their memory of a single incident to make this judgment. However, eyewitnesses may confuse unrelated incidents that happen quickly one after another. This is known as "overtaking" and it can lead to false memories. For example, if someone sees someone else commit a crime and then watches as that person is arrested by police officers, they might believe that person was responsible for the first incident even though nothing about the arrest would have been apparent to them at the time of the original sighting.
It is also possible for eyewitnesses to see something different from what actually happened. This is called "misidentification".
According to research, eyewitness identification testimony can be very unreliable. Although witnesses are frequently confident that their memory is correct when identifying a suspect, the malleability of human memory and visual perception makes eyewitness testimony one of the most untrustworthy forms of evidence. In fact, studies have shown that as many as 90% of eyewitnesses incorrectly identify suspects.
The problem with eyewitness testimony is twofold: first, the quality of witness memories decreases over time; second, the way in which witnesses describe what they see may influence judges to find guilt even if there was reasonable doubt before trial.
Research has shown that the accuracy of eyewitness identifications drops off sharply within three days of the incident. This is because the more time that passes between witnessing the event and being asked to identify its perpetrators, the less reliable our memories become.
Even if eyewitnesses do identify the right person, this does not guarantee that they will be believed by jurors. Studies have shown that the likelihood of an innocent person being convicted depends on several factors, such as the number of witnesses offered, the level of certainty expressed by each witness, and the degree to which these witnesses corroborate each other's statements. Jurors are likely to believe that honest mistakes were made during jury selection or evidence testing, especially if there are multiple suspects who resemble each other closely or if the case involves a crime of passion or revenge.