Evidence in support of this theory is provided by
Allport who showed two men (one white and one black) arguing on a subway train. Participants were more likely to remember the open razor as being in the hand of the black man, whereas it was being held by the white man. This provides evidence that stereotypes affect the accuracy of EWT as participants had a simplistic schema about a particular group of people and this affected their recall of the event.
It is not clear if schemas alter our initial perceptions or if they alter subsequent recall. Bartlett assumed it was the retrieval process that was affected, but recent research has shown that initial understanding and storage are affected. Evidence to support this is provided by Loftus and Palmer’s 2nd experiment whereby post event information in the form of leading questions (Did you see the “broken glass?”) changed the original memory for the event rather than creating a response bias. However the validity of studies such as this has to be questioned because they compromise the extent to which they represent the behaviour of real witnesses as the “crime” scenes are less realistic. This means they lack ecological validity and the results are difficult to generalise beyond the experimental setting to real life examples of eye witness testimony.
During an incident
Cognitive factors such as attributional bias may affect EWT. For example, an eyewitness may tend to overestimate dispositional (internal) factors tending to assume that a person is “criminal” in nature rather than being a consequence of circumstances (external/situational) E.g. assuming a customer in a clothes shop who trigger the alarm has shoplifted as opposed to the sales assistant leaving the tag on by mistake. This has been demonstrated by many pieces of research such as Barjonet (1980) who found that people tended to believe car accidents were caused by driver error (dispositional attribution) rather than situational factors, like driving conditions.
After the incident
Research has shown juries put more trust in eyewitness testimony than they do on forensic evidence (Loftus, 1986), which causes a problem when eyewitness testimonies can be so unreliable. There is a huge wealth of evidence demonstrating eyewitness accounts are inaccurate however we must not forget that the ability of recall will be unique person to person and some individuals will be able to present an entirely accurate account of a crime they have seen.