Memory case studies

Method: Lab experiment-Participants shown a grid with 3 rows of 4 letters for 50 milliseconds (0.05 seconds). They then had to immediately recall either the whole grid or a randomly chosen row indicated by a tone played straight after the grid was shown.

Results: On average participants could only recall 4 to 5 letters of the whole grid, could recall 3 letters on any row.

Conclusion: Can be concluded that the participants could recall any 3 from any row since they didn’t know which row they would be asked to remember. They couldn’t report the whole grid because it had faded before they could finish recall.

Evaluation: Lab experiment means highly scientific. Variables could be controlled so easily replicated. Artificial setting means lack of ecological validity.

Method: Participants were shown random trigrams (3 random consonants) and asked to recall them after 3, 6, 9, 12, 15 or 18 seconds, during the pause they were asked to count backwards in threes from a given number. This was an interference task.

Results: After 3 seconds, participants could recall about 80% of trigrams correctly. After 18 seconds, around 10% was recalled correctly

Conclusion: When rehearsal is prevented, very little can stay in STM for longer than 18 seconds.

Evaluation: The results are likely to be reliable since it’s a lab experiment but lacks ecological validity, only one stimulus was used so duration could depend on stimulus. Could have been confusion since participants saw more than one trigram

Method: 392 people were asked to list the names of their ex-classmates (free-recall test). Shown photos and asked to recall the names of the people shown (photo-recognition test) or given names and asked to match them to photos of classmates (name-recognition test).

Results: Within 15 years of leaving school could recognise 90% of the names and faces. They were 60% accurate on free recall. After 30 years, free recall had declined to about 30% accuracy. After 48 years, name-recognition was about 80% accurate, and photo-recognition about 40% accurate.

Conclusion: Evidence of VLTMs is a “real-life” setting. Recognition is better than recall, so there may be a huge store of information, but it’s not always easy to access all of it-you just need help to get to it.

Evaluation: This was a field experiment so has high ecological validity but hard to control all variables, making these variables less reliable-no way of knowing exactly why info was recalled well. It showed better recall than other studies on LTM, but this may be because meaningful info is stored better. This type of info could be rehearsed if still in contact with classmates which increases the rate of recall. Meaning the results can’t be generalised to other types of info held in LTM.

Method:Participants shown string of letters or digits. Had to repeat them back in the same order. Number increased until the participant failed

Results: Participants recalled around 9 digits and 7 letters. Capacity increased with age during childhood.

Conclusion: STM has a limited storage capacity of 5-9 items. STM increased with age possibly due to chunking. Digits may be easier to recall as only 10 of them.

Evaluation: Artificial so lacks ecological validity. Previous sequences may confuse them on future trials.  

Method: Participants were given four sets of words that were either acoustically similar (e.g. man, mad, mat), acoustically dissimilar (e.g. pit, cow, bar), semantically similar (e.g. big, large, huge) or semantically dissimilar (e.g. good, hot, pig). The experiment used an independent groups design- participants were asked to recall the words either immediately or following a 20 minute task.

Results: Participants had problems recalling acoustically similar words when recalling the word list immediately (from STM). If recalling after an interval (from LTM), they had problems with semantically similar words.

Conclusion: The patterns of confusion between similar words suggests that LTM is more likely to rely on semantic coding and STM on acoustic coding.

Evaluation: Lacks ecological validity. Also, there are other types of LTM (e.g. episodic, memory, procedures) and other methods of coding (e.g. visual) which isn’t considered in this experiment. This experiment used an independent groups design, so there wasn’t any control over participant variables.

Method 1: Participants were shown a film of a car crash. They were asked questions including “How fast do you think the cars were going when they hit” In other conditions, the word “hit” was replaced with “smashed”, “collided”, “bumped” or “contacted”.

Results: Participants given the word “smashed” estimated the highest speed (41 mph); those given the word “contacted” gave the lowest estimate (32 mph).

Method 2: Participants were split into 3 groups. One group was given the verb “smashed”, another “hit”, and the third group wasn’t given any indication of the vehicle's’ speed. A week later, participants were asked “Did you see any broken glass?”.

Results: Although there was no broken glass in the film, participants were likely to say that they’d seen broken glass in the “smashed” condition.

Conclusion: Leading questions can affect the accuracy of people’s memories.

Evaluation: This has implications for questions in police interviews. As an artificial experiment this wasn’t emotionally arousing which may affect recall. Experimental design might lead to demand characteristics, where results are skewed because of participants’ expectations about the purposes of the experiment. This would have reduced the validity and reliability of the experiment.

Method: In a study with an independent groups design, participants heard a discussion in a nearby room. In one condition, a man came out of the room with a pen and grease on his hands. In the second condition the man came out carrying a knife covered in blood. Participants were asked to identify the man from 50 photos.

Results: Condition 1 was 49% accurate but only 33% in condition 2.

Conclusion: When anxious, witnesses focus on a weapon at the expense of other details.

Evaluation: High ecological validity as the participants didn’t know they were being studied but there are ethical considerations as participants could have been distressed at the sight of the knife.

Method: Participants were shown a film of a car crash. They were then asked either “Did you see the broken headlight?” or “Did you see a broken headlight?”. There was no broken headlight.

Results: 17% of those asked about “the” broken headlight claimed they saw one, compared to 7% asked about “a” broken headlight.

Conclusion: The use of the word “the” is enough to affect the accuracy of memories.

Evaluation: Lab study allowed control over variables so it’s possible to establish cause and effect but lacked ecological validity.

Method: 3 groups of participants (children, young adults and the elderly) watched a video of a kidnapping. They were then asked a series of leading and non-leading questions.

Results: The elderly and children gave more incorrect answers to non-leading questions. Children were misled more by leading questions than the elderly.

Conclusion: Age has an effect on the accuracy of eyewitness testimony.

Evaluation: Implications in law when children and elderly are involved. But as an artificial experiment it wasn’t emotionally arousing which could affect recall. It lacks ecological validity.

Method: In a staged situation, an intruder carrying a blue rucksack entered a classroom and stole a slide projector. 2 days later, participants were questioned about the event. The study used an independent groups design-participants were either questioned using a standard interview procedure or the cognitive interview technique. Early in the questioning, participants were asked “Was the guy with the green backpack nervous?”. Later in the interview, participants were asked what colour the backpack was.

Results: Participants in the cognitive interview condition were less likely to recall the rucksack being green than those in the standard condition.

Conclusion: The cognitive interview technique reduces the effect of leading questions.

Evaluation: High ecological validity. Participants in the cognitive interview could be less susceptible to leading questions due to individual differences.

Method: Each participant was given either 1, 2, 3, 4, 5 or 6 lists of 24 words. Each list was divided into 6 categories of 4 words. Words were presented in category order. After the lists were presented, in one condition, participants had to recall all the words (free recall). In another condition participants were given a category and had to remember all those words (free cued recall). 

Results: In free recall there was evidence of retroactive interference. In free cued recall retroactive interference disappeared. 

Conclusion: Suggests interference doesn't cause forgetting. The forgetting shown was cue-dependent. 

Evaluation: Lab experiment so can't be generalised as it lacks ecological validity. 

Underwood and Postman carried out a study which supported retroactive interference. In a lab experiment, participants were split into two groups. Both groups were given a list of paired words to learn, e.g. cat-tree. The experimental group was then given a second list of words to learn, where the first of the words in each pair was the same as in the first list, e.g. cat-dirt. The control group wasn't given a second word list. Both groups were then tested on their recall of the first word list, by being given the first word from each pair. Recall was better in the control group, suggesting that retroactive interference of the second word list had affected recall for the experimental group.  

Underwood studied proactive interference by looking at the results of studies into forgetting over a 24-hour period. He found that if people had previously learnt 15 or more word lists during the same experiment, a day later their recall of the last word list was around 20%. If they hadn't learnt any earlier lists, recall a day later was around 80%. Underwood concluded that proactive interference from the earlier lists had affected the participants' ability to remember later ones.

Shaw et al paired participants with a confederate (who pretended to be another participant). The pairs were shown videos of a staged robbery and were interviewed together afterwards. The participant and the confederate alternated who answered the questions first. When the participant was first, recall was accurate 58% of the time. When the confederate answered first and gave the correct answer, the recall of the participants was 67%. If the confederate gave the wrong answer, correct recall of the participant fell to 42%.

Gabbert et al's study involved 2 groups of participants- young adults (27-33 years old) and older adults (58-80 years old). Both groups watched a staged crime and were exposed to misleading information in one of two ways- through conversation with a confederate who was pretending to be another participant, or reading a written report of the crime, supposedly written by another participant. The participants were then given a recall test about the event they'd witnessed. It was found that both groups of adults were more likely to report inaccurate information after a conversation with a confederate than after reading the report. 

Carried out a case study into a patient called HM who had suffered from severe and frequent epilepsy. His seizures were based in a brain structure called the hippocampus. Doctors decided to surgically remove part of the brain around this area. The operation reduced his epilepsy, but led to him suffering memory loss. He could still form short term memories, but was unable to form new long term memories. This supports the theory that different types of memory are seperate systems in the brain. 

KF was a brain-damaged patient who had an impaired STM. His problem was with immediate recall of words presented verbally, but not with visual information. This suggested he had an impaired articulatory loop but an intact visuo-spatial sketchpad, therefore providing evidence for the working memory model's view of STM. This finding couldn't have been explained using the multi-store model of memory, which proposed that short-term memory was just one system. 

Participants were split into two groups. All of the participants had to carry out a task where they had to follow a moving spot of light. This would use the visuo-spatial sketchpad. At the same time, one group of participants also had to describe the angles on a letter- another task involving the visuo-spatial sketchpad. The other group of participants were given a second task that would use the phonological loop- they had to do a verbal task whilst following the light. Gathercole and Baddeley found that performance was much better in participants doing tasks that required seperate systems.