Bartlett says that humans make sense of information by putting them into schemas—units of memory that correspond to different situations. Schemas allow people to understand what they encounter. They may be determined by one’s prejudice—Bartlett’s Reconstructive Memory theory suggests that recall is subject to interpretation, based on a person’s norms and values. This means that we recall memories based on what we would expect the normal thing to be is.
He tested this is using different stories—he used the story called “the War of the Ghosts” on English participants (who were unlikely to have heard the Native American story). He found that after the story had been reproduced through six people that it had changed in some ways: it was usually half of the original length and the details that were of the Native American culture were often left out, this made the story reflect the participants own culture as opposed to the story’s.
This study brings us to the conclusion that people reconstruct a memory to fit their own culture, and what they think is normal.
This shows that people’s memories are not a useful tool in eyewitness testimony—as details key to the trial may not be recalled because they do not fit the person’s view of what is normal, and therefore is not a part of the reconstructed memory.
A study by Allport & Postman also showed how stereotypes can affect a person’s memory. The study consisted of asking participants to recall a picture they had been shown of a smart-looking black man being attacked by a rough-looking white man with a razor; participants tended to recall the picture as the black man attacking the white one with the razor. This shows that memory is reconstructed by what we perceive to be the normal, stereotypical thing.
If this is applied to eyewitness testimony, it would result in the victim of a crime being prosecuted as opposed to the actual perpetrator.
There are other theories as to why memories may not be accurate and why that means eyewitness testimonies are unreliable. Levels of Processing theory explains how there are different levels of encoding (forming a memory): shallow, intermediate and deep, corresponding to structural, phonetic and semantic respectively. This suggests that memory may become inaccurate because of the way it is originally processed—Craik and Tulving found that if it is processed through a shallow level of encoding, it is unlikely to be recalled and especially not with accuracy; however, conversely, a deep level of encoding will make it more likely for a memory to be recalled with accuracy. They also found that semantic encoding is the best way of forming a strong, reliable memory.
The two theories differ because Levels of Processing goes further to explain why certain memories may not be as strong as others, based on the level of which a memory is formed; whereas Reconstructive Memory theory suggests that it is to do with a person’s opinion and attitude on what a “normal” thing to see is, and is manipulated from that opinion. However, both studies offer insight as to why eyewitness testimonies may become unreliable as evidence in a court of law.
Reconstructive Memory theory, therefore, suggests that as a result of memories being created and stored through schemas and imaginative reconstruction, memories become unreliable as a piece of evidence because people piece the memory back together according to what they think should be the normal way the memory went. This explains how eyewitness testimony is unreliable because it means that despite what actually happened, an eyewitness may give an account leading to the prosecution of a wrongly accused person because they have reconstructed the memory in an incorrect way as a result of stereotyping and their own beliefs.