The Memory Process and Reconstructive Memories

Memory is usually divided into three main processes: encoding, storage, and retrieval. To understand why witnesses sometimes give incorrect or incomplete accounts, it is necessary to understand how these three processes work.

• Encoding is the process by which a memory is created as we experience an event.
• Storage is the period between the event and the point at which the memory is to be recalled.
• Retrieval is the process by which we try to bring forward and describe what we previously experienced.

Each of these processes can be affected by different factors, which means memory can change before it is recalled.

Encoding — when memories are created

Encoding means that information is registered and converted into a memory. How well a memory is encoded greatly affects how detailed and accurate it can later be recalled.

For witnesses, encoding often takes place under highly unfavorable conditions. The event may be brief, unexpected, and emotionally intense. In addition, visibility may be limited or attention may be directed at only certain parts of the event. This can make the memory incomplete from the very start.

The importance of exposure time

How long a witness observes an event greatly affects how much information there is time to encode. If a person only sees an offender for a few seconds, the ability to later describe that person becomes considerably poorer than if the observation lasts several minutes.

At the same time, a longer exposure time does not mean the witness automatically remembers everything that happened. A person who, for example, is kidnapped over several days cannot be expected to remember every detail. Memory capacity is limited, and even long events contain more information than the brain can register and store.

Emotionally charged situations

One of the most important factors in encoding is how emotionally charged an event is. Research shows that people who experience very strong emotions often become better at remembering central information, that is, the information that matters most for the course of events, such as the offender's actions or the threat to the victim. However, they often become worse at remembering peripheral information, that is, details of lesser importance to what is happening. Examples of peripheral information include what music was playing in the background or which people were standing further away.

The American psychologist Donald Easterbrook presented a theory usually called Easterbrook's hypothesis. The hypothesis states that strong stress or fear causes a person's attention to concentrate on what is perceived as most significant. Attention therefore becomes more limited, a phenomenon often described as tunnel vision.

It is important, however, to distinguish between what the witness actually registers and what the person can later recall. Modern research shows that tunnel vision appears to be mainly mental rather than purely visual. The witness often registers more information than they later remember directly after the event. Once the acute stress has settled and the person has gained emotional distance, even previously inaccessible details can sometimes be recalled.

The weapon focus effect

When a weapon is present during a crime, a large part of the witness's attention is directed at the weapon. This phenomenon is called the weapon focus effect. Because the weapon is perceived as the greatest threat, other details receive less attention. The witness may therefore have difficulty describing, for example, the offender's face, clothing, or other identifying features.

This means a witness may well give a detailed description of the weapon while the description of the offender is considerably less reliable.

Strong emotions affect not only what we remember but also how we perceive time. During traumatic events, people tend to overestimate how long a situation lasted. An event that in reality lasted about a minute can be experienced as considerably longer.

This altered perception of time does not mean the witness deliberately gives incorrect information. It is a natural consequence of how the brain processes highly stressful situations.

Storage and forgetting

Storage is affected by how much time passes between encoding and the point at which we are to recall the information. We forget things. Usually, a large part of the memory disappears immediately after the event, and after that the memory continues to deteriorate, but at a considerably slower rate.

Ebbinghaus carried out some of the earliest scientific experiments on memory. In these, he had participants memorize nonsense words made up of three letters, for example GUK, VUL, NOV, and so on. Participants were given a list of these nonsense words to learn and were then asked to recall the words. Once they had learned the list, he let a certain amount of time pass, after which they again tried to learn the same list.

In the experiment, he noted that already after one hour following the learning session, participants needed to spend at least 50% of the original learning time to relearn the words and reach the same performance. In other words: if it took 10 minutes to learn the words the first time, at least 5 minutes of repetition was required an hour later to achieve the same result.

The experiment also showed that the order of the nonsense words mattered. Participants remembered the words at the beginning of the list and those at the end better. This is likely because the words at the beginning of the list have more time to be processed in working memory, which increases the chance that they are stored in long-term memory. This is called the serial position effect.

Participants remembered the words at the end of the list better than those in the middle if the test took place immediately after learning. If, however, they were first distracted before the test, this effect disappeared. This is likely because the words are still present in working memory during an immediate test but disappear with distraction. This is called the recency effect.

Both the serial position effect and the recency effect can affect what we remember in connection with a trial. Information presented first, and also information presented last, for example in a summary, is often remembered better. This means the opening and closing of a trial can become particularly important for its outcome.

Retrieval and reconstructive memories

Our memories are reconstructive, which means they change over time. Every time we think about a memory, we create a new version of it. Information can be added, and other information can disappear.

Interference means that earlier or new information becomes mixed up with the memory we are trying to recall. For example, I might mix up the memory of Christmas two years ago with both earlier and later Christmases. The longer the storage period, the greater the risk of interference.

An experiment showing how reconstructive memories can be affected by new information might look like this: a group of 40 people watch a filmed reenactment of a crime in which the perpetrator wears a gray t-shirt. The group is then split into two halves.

• 20 people read an accurate summary of the film.
• 20 people read a summary containing a false claim, namely that the t-shirt was green.

Afterward, everyone answers questions about the film, including: "What color was the perpetrator's t-shirt?"

Those who received the misleading information often give the wrong answer and believe the t-shirt was green. The reconstruction of the memory can sometimes create an entirely new memory, but sometimes a blended memory arises in which the original information and the new information are combined.

Another experiment showing this involved a film in which eight loud demonstrators disrupt a professor's lecture.

• Half of the participants were asked: "Was the leader of the four demonstrators a man?"
• The other half were asked: "Was the leader of the twelve demonstrators a man?"

When participants were asked a week later to recall what they had seen, those who had been asked about twelve demonstrators estimated they had seen an average of 8.9 people, while the other group estimated the number at 6.4 people.

Research shows, however, that changes to memories tend to occur to a greater extent in connection with less essential information, and to a lesser extent in connection with the more central parts of a sequence of events.