AS Psychology Key Studies Memory Memory: what you need to know • The multi-store model of memory: sensory register, short-term memory and long-term memory. • Features of each store: coding, capacity and duration. • Types of long-term memory: episodic, semantic, procedural. • The working memory model: central executive, phonological loop, visuo-spatial sketchpad and episodic buffer. Features of the model: coding and capacity. • Explanations for forgetting: proactive and retroactive interference and retrieval failure due to • absence of cues. • Factors affecting the accuracy of eyewitness testimony: misleading information, including leading • questions and post-event discussion; anxiety. • Improving the accuracy of eyewitness testimony, including the use of the cognitive interview. Key Term Meaning Encoding/Coding Changing the format of information for use in memory Capacity The amount of information that can be held in memory Duration The length of time information remains in memory Sensory register A store of sensory information that lasts no more than a few seconds Displacement/displaced Information that was being held is pushed out by new information that enters our memory What is memory? The term memory can refer to the process in which we retain information and knowledge. You may think that this is one process but in fact it is split in to 3 basic processes. The first process is called Encoding. This is where a sensory input (e.g. sounds or images) are transformed for it to be registered in memory. Sound, vision and meaning are all encoded differently. The second process is Storage. This simply is holding information until it is required. The third stage is the process of Retrieval. This is where we locate and use the information we have stored so we are consciously aware of it. Short Term Memory (STM) and Long Term Memory (LTM) Psychologists believe that Short Term Memory and Long Term Memory differ in the amount of information that can be stored, encoded and the duration the information can be held. Short Term Memory (STM) is a system for information that you are currently thinking of. Information that is stored here is done so only on a temporary basis. Encoding in STM: Psychologists believe that information in STM is mainly encoded acoustically (sound). Capacity of STM: The capacity (how much of something can be held) is limited in STM. Only a small amount of information can be held. This can be demonstrated through studies using Serial Digit Span techniques. Table of differences between STM and LTM: ENCODING STM LTM CAPACITY DURATION Atkinson and Shiffrin’s Multi-Store Model of Memory Sensory Memory Iconic register: ‘Icon’ is another word for image or picture. The iconic register therefore refers to our memory for visual information. Sperling (1960): Iconic Register Atkinson and Shiffrin based their assumptions about iconic memory on the findings of previous research. In particular, that of Sperling. In his experiment, Sperling very briefly displayed to his Ps visual arrays containing three rows of letters: M N J C N X K V F L B P He found that Ps could only recall four or five letters from the 50 millisecond arrays, but reported being aware of more letters that they could not report. Sperling assumed that, because visual information was only available for a very short time, it faded in memory before Ps could recall it. Sperling conducted a further experiment to investigate this, using what he called a partial report procedure. He trained Ps to recognise 3 tones – high, medium and low. The high tone was related to the top row of letters, the middle to the middle row of letters and the low to the bottom row of letters. Sperling instructed Ps to learn which tone matched with which row of letters. One Ps had learnt this, Sperling once again presented Ps with a series of digits for 50 milliseconds each. This time along with a tone, cueing Ps to a particular row. He found that Ps recalled on average 75% of the letters in the cued rows, a much better % than the whole array. Sperling concluded that Ps performance improved because a row contained fewer items than the whole display, therefore there was less decay of information from memory before Ps had to recall it. Recall was not 100% however, Sperling saw this as evidence that not only does the iconic register have a very limited capacity but that information decays and is lost very rapidly. Echoic register: the sensory register for auditory information. Darwen et al. (1972) suggest that the length of time information is stored in echoic memory is about 3 seconds. Darwen et al. (1972) Ps were presented with spoken recordings of letter and number lists. The lists were presented over headphones so that it seemed that one list came from the left and one list came from the right and a third from behind. After hearing the lists, Ps were given a cue to recall one of the three lists. The length of time from the presentation to the cue varied from 0 – 4 seconds. The cue should have made the task easier as then Ps only had to recall a particular list from a particular direction. Darwen et al. found that as the time between presentation and cue increased, the recall performance of Ps decreased. After a 3 second delay Ps performed no better than they would have without cues. Further testing showed that performance did not significantly improve when Ps were cued to recall letter or digit lists. Capacity of STM Jacobs (1887): Capacity of STM Jacobs read allowed a list of either letters or numbers, with one syllable (‘w’ and ‘7’ were excluded). The lists increased until participants could recall them only 50% of the time. A wide age range was used in his sample. Jacobs found that STM capacity for digits was 9, whereas letters was 7. The conclusion was that STM has a limited capacity of 5-9 digit and that numbers were easier to recall than letters. Criticisms show that practice can affect information recall and also newer research shows people have differing abilities associated with their attention levels. Capacity of STM – Miller (1956): ‘Chunking’ George Miller (1956) made observations of everyday practice. For example, he noted that things come in sevens (7 notes on the musical scale, 7 days of the week etc.). This suggests that the span of the STM is about 7 items 9plus or minus 2). However, Miller also noted that people can recall 5 words as well as recall 5 letters. They do this by chunking – grouping sets of letters or digits together into units or chunks. Limitations of Miller: Cowan (2001) who reviewed other research and concluded that the capacity of STM was only about 4 chunks. Simon (1974) found memory span as measured in chunks depends on the amount of information contained in the chunk. He investigated the ability of Ps to recall one-syllable, two-syllable and three-syllable words and two-word and eight-word phrases. He found that larger chunks resulted in reduced memory span, so that Ps accurately recalled fewer larger chunks (e.g. eight-word phrases) and more smaller chunks (e.g. two-word phrases). Duration of STM The duration of STM is short - less than 30 seconds. This is best evidenced by Peterson and Peterson (1959) who used nonsense trigrams (3 letters with no meaning, i.e. WQT). Ps were instructed to read the trigrams and then straight after had to count backwards in 3’s from a large number (e.g. 375). They counted for a specific time period. The participants had specific times in which to recall the trigrams ranging from 3 seconds to 18 seconds. (The counting was done to stop the retention of the trigrams.) The findings showed that 90% of trigrams were recalled after a 3 second interval and only 6% after an 18 second interval proving that the duration of STM is a matter of seconds. One criticism of Peterson and Peterson’s study is that the stimulus for memory was artificial. Trying to remember consonant syllables does not reflect real-life memory activities, so we might say that this study lacks external validity. Craig and Tulving (1973) stated it is not the amount of rehearsal time in STM that determines long-term retention, rather, it is how the information is rehearsed. They suggested that there are two types of rehearsal – maintenance rehearsal, which is when a person continually repeats the information and elaborative rehearsal – which is where the information is transformed in to something meaningful. They stated that only elaborative rehearsal allows information to transfer in to LTM, for example, you will remember a phone number that is similar to another phone number you know – you will attach meaning to it by linking it to the previous phone number. Sebrechts et al. (1989) Demonstrated the idea that duration of STM depends on a number of factors. In addition to rehearsal, we can extend duration by our intention to recall the information later. They tested recall for sets of three familiar English nouns. In the condition where Ps were not expecting to be asked to remember the words, correct recall fell to 1% after only about 4 seconds. Encoding in STM and LTM: Most psychologists believe that information is encoded semantically (meaning) in LTM and acoustically (by sound) in STM. Baddeley (1966) illustrated this in a study which involved lists of words. List A had acoustically similar words (cat, mat, sad, sat). List B had acoustically dissimilar words such as (pit, day, cow). List C had semantically similar words such as (big, huge, tall). List D had semantically dissimilar words such as (hot, safe, foul). After they were given the list they had a 20 minute retention interval where they performed another task. This ensured that recall would have to come from LTM. The list C (semantically similar) was the least well remembered (55%). This seems to suggest there was semantic confusion and this leads to the idea that LTM information is encoded semantically. The support for this comes from cognitive sense. In everyday life we can remember the meaning of things but not the words. An example would be a TV programme, you can remember what happened (semantic) but not the exact words (acoustic). Baddely also repeated this experiment but instead of leaving a 20 minute interval he asked Ps to recall immediately after seeing them. He noted that Ps tended to perform less well on the list with acoustically similar words. He therefore concluded that STM encodes acoustically. Conrad (1964) – encoding in STM as acoustic Showed Ps a random sequence of consonants, projected very rapidly on to a screen. Some Ps were shown consonants that were acoustically similar, that is, they sounded similar when spoken (B, C, T, D,). Other Ps were shown consonants that were acoustically dissimilar (R, E, J) Immediately following the presentation, Ps were asked to write the letters down in order that they were shown. Since this task was well within the 5-9 capacity of STM it should have been relatively easy for Ps. However, Conrad found that Ps frequently made errors of recall. Ps found it more difficult to accurately recall sequences that sounded the same compared with sequences that sounded different. Conrad concluded that while the consonants had been presented visually, they had been encoded acoustically by Ps in STM. It was this encoding that caused difficulties in recalling letters that sounded similar. Limitations of STM encoding acoustically: Shulman (1970) presented Ps with a series of word pairs and required them to indicate as fast as possible if one of the two words (the ‘probe’ word) was identical to the other, a homonym of the other (prey-pray) or a synonym of the other (leap-jump). He found a similar performance for all three probe types, leading him to conclude that semantic encoding is possible in STM as Ps had obviously had to understand the meaning of the words in order to complete the task. Capacity and Duration of LTM Capacity and Duration of LTM The potential capacity of LTM is unlimited. This is because we know that our brains are not full of information on everything. This is why there has been no limit placed on it. In LTM the duration lasts as long as you live. You will have memories that you have from childhood that will last until you die. Information in LTM doesn’t have to be rehearsed or repeated. For example you may not have ridden a bike for many years but you can still do it. Bahrick et al. (1975): Duration of LTM Duration in LTM was demonstrated in a study by Bahrick et al (1975). They gathered 400 participants aged 17-74 years and asked them to remember names of classmates from high school (free recall task). They were also shown a list of names and photos. They had to identify their ex-schoolmates (recognition task). Those who left school within the previous 15 years recalled 90% of the names and faces in the recognition task. Those who had left within the previous 48 years recalled 80% of names and 70% of faces. These high percentages show we remember names and faces for a very long time. The amount of time information stays in LTM is also determined by how well the information was learned in the first place. Bahrick and Hall (1991) Tested LTM for geometry and algebra. People who had only taken maths courses up to secondary school level showed a steady decline in their recall accuracy over the years. However, students who had gone on to take a higher level course in maths showed greater levels of accurate recall as much as 55 years later. See encoding in LTM and STM page for evidence that LTM encodes semantically Strengths and weaknesses of the MultiStore Model of Memory Strengths: One of the greatest strengths of Atkinson and Shiffrin’s assertion that STM and LTM are different memory stores comes from people that have suffered brain damage. The loss of memory among such people is usually selective – that is, it affects one type of memory but not another. Shallice & Warrington (1970) The case of KF KF was a young man who sustained brain injuries after a motor cycle accident. He had an impaired STM working alongside a fully functioning LTM. He appeared to have an intact LTM in that he was able to learn new information and recall stored information. However, his STM had a much reduced capacity so that he was only able to store a couple of bits or chunks of information rather than the normal 5-9 chunks. Drachman and Sahakian (1979) Alzheimer’s disease is a serious disorder of the brain, and early symptoms include severe memory impairment. Researchers have been interested in investigating some of the specialised chemicals in the brain called neurotransmitters , which are involved in brain processes. Patients with Alzheimer’s disease have been found to have low levels of acetylcholine, compared to people without the disease. D & S investigated this by administering to a group of Ps a drug that clocks the action of acetylcholine in the brain . They then gave Ps tasks that tested either their LTM or STM and compared their performance with Ps that had not had the chemical blocked. They found that Ps that had had the chemical blocked performed at normal levels for their STM but more poorly on the LTM tasks. This suggests that STM and LTM work as separate stores involving different neurotransmitters. Squire et al. (1993) found during brain scans that different parts of people’s brain were activated for STM and LTM tasks, suggesting again that STM and LTM are separate stores (hippocampus more active during LTM, prefrontal cortex more active during STM). Weaknesses: The MSM has been accused of being overly simplistic – human memory is extremely complex, and it is highly unlikely that such a simple model could reflect this complexity. Brown and Kulik (1977): ‘Flashbulb memory’ Flashbulb memory is a special type of remembering that involves the insignificant details surrounding an emotional or shocking event being imprinted on to long term memory without rehearsal. For example, those involved in the terrorist attacks in Paris will not have needed to rehearse the finer details of the event in order to remember them, they will be imprinted onto their long term memory automatically. Ruchkin et al. (1999) Measured brain activity in Ps who were required to listen to a set of words and pseudo-words. If people only process information acoustically in STM, there should be no difference in brain activity when processing words and pseudo-words. However, Ruchkin et al. found that there were considerable difference in the recall of the two types of word, which suggests that semantic information stored in LTM was being used in this task. Brandimonte et al. (1992) Prevented acoustic coding in STM by asking Ps to learn a list of words whilst reciting a meaningless chant (La, La, La). Found that Ps accuracy of recall was not affected and concluded that STM can use visual coding as opposed to acoustic and went so far as to say that visual coding is actually more effective than acoustic. Working Memory Model Strengths: the WMM has been extremely influential and most cognitive psychologists now use the term working memory in preference to the term STM. It is a much more plausible model than the MSM because it explains STM in terms of both temporary storage and active processing. It also incorporates verbal rehearsal as just one optional process within the articulatory loop, instead of being the sole mode and means of transferring information to LTM, as suggested by Atkinson and Shiffrin. Baddely et al. (1998) Presented evidence that the phonological loop plays a key role in the development of reading and that the PL is not functioning properly in some children with dyslexia. It is less important for adult readers but still has an important role in helping to understand complex text. Gathercole and Baddeley (1993) Participants followed a spot of light with a pointer (tracker task). At the same time half the participants had to describe the angles on a letter ‘F’. This group found the task very hard as they were using the visuo-spatial scratchpad for both tasks. The other group did a verbal task along with the tracker task and had little difficulty performing both tasks. Explanations for Forgetting Retroactive Interference: Waugh and Norman (1965) Presented Ps with a sequence of 16 digits, after which they were shown one of the digits from the list (the ‘probe). Ps had to say which digit in the sequence occurred just prior to the probe. The researchers found that the fewer items there were following the probe in the original sequence, the more likely Ps were to recall the preceding digit correctly. This demonstrates retroactive interference. If the probe occurred early in the sequence then the greater numbers of digits following it increased the chances of interference. Likewise, if the probe occurred towards the end of the sequence then there would be fewer digits to cause interference. Proactive Interference: Keppel and Underwood (1962) Ps required to recall consonant trigrams after varying intervals, during which they counted backwards in 3s. While forgetting was found to increase with the interval, Keppel and Underwood noted that there was little or no forgetting of trigrams from the start of the procedure. They concluded that the earlier trigrams had entered into LTM and were interfering with the memory for later trigrams. Retrieval Failure Tulving and Thompson (1973): Encoding-specificity principle More likely to forget something when the context during the encoding is absent when you try to retrieve it. For example, when you decide in one room of the house you want something from another area of the house, once you get to the next room in the house you have forgotten what you went in there for. Tulving and Thompson would argue that this is because in the next room, the cues that were there when you encoded what you needed to remember are absent in a different room. Marian and Fausey (1986) Context-dependent forgetting – relevant environmental cues are absent during recall Found that memory for a story was better if the language in which it was presented and the language that was used to test memory were the same. Ps in their study were residents of Chile who were fluent in their native language (Spanish) and English. They were required to listen to four academic type stories, two of them spoken in Spanish, two of them spoken in English. Shortly afterwards they were asked to recall features of each story and answer in the same language as the question. Half the questions matched the language of the story (e.g. English story – English questions) and half did not (English story – Spanish questions). M & F found that Ps who heard the story and questions in the same language had greater accuracy of recall than those who had heard the story and questions in different languages. This demonstrated encoding specificity, as memory is better when there is a match between context during encoding and retrieval – the language used provided a retrieval cue for the recall of story details. Peters and McGee (1982) State-dependent forgetting - relevant physiological or psychological cues that were present when encoding are absent during recall Gave half their Ps a low nicotine content cigarette before learning a list of nouns. Ps who were in the same state the next day (having smoked or not smoked a cigarette) performed better on a recognition task than those who were in a different state. Peters and McGee concluded that the arousal caused by the cigarette had helped cause a physiological context, creating a state-dependent effect. Eye-Witness Testimony Misleading Information Research tells us that one of the main factors affecting the accuracy of memory for an event seems to be what happens after the event has taken place. Misinformation after an event can introduce serious errors into eyewitnesses’ recall of the event. Loftus (1992) calls this ‘misinformation acceptance’, where people accept misleading information after an event and absorb it into their memory for the actual event. Loftus and Palmer (1974): Leading Questions A leading question is one that by its content or format has the answer embedded within it. This has been found to influence the eye witness account of an event. This was studied by Loftus and Palmer. The procedures involved 45 students being shown 7 films of different traffic accidents. After each film the participants were given a questionnaire which asked them to describe the accident and then answer a series of specific questions about it. There was one critical question: “About how fast were the cars going when they hit each other?”. One group of participants was given this question. The other four groups were given the verbs ‘smashed’, ‘collided’, ‘bumped’ or ‘contacted’ in place of the word ‘hit’. The mean speed estimate was calculated for each group, as shown in the table. The group given the word ‘smashed’ estimated a higher speed than the other groups (about 41mph). The group given the word ‘contacted’ estimated the lowest speed (about 32mph). Loftus and Palmer concluded that the form of question can have a significant effect on a witness’s answer. In other words, leading questions can affect the accuracy of memory. Such leading questions are an example of what psychologists call post-event information – information given after the event which may alter memory. It is possible that such post-event information causes the information to be altered before it is stored so that memory is permanently affected. A second possible explanation is that the form of the question actually alters the participant’s memory representation of the accident, which leads them to produce a higher or lower estimate. Loftus and Zanni (1975): Leading Questions Showed Ps brief film clips of a car accident and then asked a series of questions. Half the Ps were asked if they had seen ‘a’ broken headlight, the other half were asked if they had seen ‘the’ broken headlight. Although there was no broken headlight in the film, 17% of people asked about ‘the’ broken headlight reported seeing one as opposed to only 7% reporting seeing on in the ‘a’ broken headlight group. Using ‘the’ had implanted the idea that there was indeed broken glass, leading some Ps to change their recall accordingly. Gabbert et al. (2003): Post Event Discussion Studied Ps in pairs. Each P watched the same video of a crime being committed, but filmed from different points of view. This meant that each P could see elements in the event that others could not. For example, only one P could see the title of a book being carried by a young woman. Both Ps then discussed what they had seen before, individually completing a test of recall. Gabbert et al. found that 71% of the Ps mistakenly recalled aspects of the event that they did not see in the video but had picked up in the discussion. Gabbert concluded that witnesses often go along with each other, either to win social approval or because they believe the other witnesses are right and they are wrong. Effects of Anxiety on EWT Deffenbacher (1983) was one of the first to investigate links between stress and EWT. He found that as we become moderately stressed/anxious, performance in EWT improves. As we hit the peak of stress our levels of accuracy drop because we feel fatigued. Loftus (1979): The Weapon Effect Ps asked to sit outside a lab where they thought they were hearing genuine exchanges between people inside the laboratory. In one condition, they heard an amicable discussion about an equipment failure. A man with greasy hands came out of the lab holding a pen. In the other condition they heard a hostile discussion, followed by the sound of breaking glass and overturned furniture. A man then emerged from the laboratory holding a knife covered in blood. Following these events, Ps were given 50 photos and asked to identify the man that had come out of the laboratory. It was found that Ps that had witnessed the peaceful scene were more accurate in recognising the man than people who had witnessed the more violent scene. Loftus believed that the anxiety caused by seeing the blood-stained knife narrowed the focus of attention for the witness and took attention away from the face of the man. From this, Loftus coined the phrase ‘The Weapon Effect’ – the eyewitness will pay particular attention to the weapon that a perpetrator uses during a crime, resulting in other aspects of the scene going unobserved. Criticisms of Loftus: Yuille and Cutshall (1986) Interview 13 witnesses to an actual crime (not one that was laboratory based as all of Loftus’ had been) some four months after the event. The researchers found that recall was very accurate, despite the inclusion of leading questions in the interview. This suggests that both misleading information and anxiety might not have the same impact on real-life witnesses as it does in laboratory-based studies. Christianson and Hubinette (1993) Carried out a survey among 110 people who had witnessed between them 22 genuine bank robberies. Some of those people had been threatened by the robbers during the event while others hadn’t. They found that those that had been directly threatened by the robbers during the robbery (so would have felt the most anxiety) actually had more detailed recall than those that hadn’t been threatened. This finding suggests that in real incidents involving high levels of stress, memory can be accurate, detailed and long-lasting. Cognitive Interview Fisher and Geiselman reviewed memory literature – people remember things better if they are provided with retrieval cues. They then developed the cognitive interview, this technique has 4 components: 1. Report everything (report every single detail of the event). 2. Context reinstatement – recreate the internal and external environment ie what mood were you in? what was the weather like? 3. Changing the order - try alternative ways through timeline of incident as it is productive to vary the different ‘routes’ into an individual’s memory. 4. Changing the perspective - imagining how it would appear from another witnesses’ point of view . Evidence Supporting the effectiveness of the Cognitive Interview Kohnken et al conducted a meta-analysis of 53 studies and found on average a 34% increase in amount of correct information generated in CI. Milne and Bull (2002) Found that when they used a combination of “report everything” and “mental reinstatement”, participants’ recall was significantly higher. This suggests the CI is relatively effective in increasing recall. Stein and Memon (2006) compared normal interviewing technique and CI in Brazil. CI increased the amount of correct information obtained and the richness (detail) of information. Weaknesses of the Cognitive Interview Kebbell et al. (1999) carried out a survey of police officers in the UK and found that there was quite widespread use of the CI. However, while officers generally found it useful, they expressed some concern about the amount of incorrect recall generated and the amount of time it took to carry out. In practice, it seemed that the officers were using the RE and CR instructions, but rarely using the CP and RO instructions. Geiselman et al (1985) found that the cognitive interview compared favourable against a standard interview. The cognitive interview has since been improved and now includes an attempt to build a greater rapport between the witness and interviewer. This is called the enhanced cognitive interview. However this means that it is difficult to evaluate the effectiveness of the cognitive interview as the use of the enhanced cognitive interview introduces a ‘lack of control’. It is difficult to draw conclusions about the effectiveness of the interview when some interviewers have better social skills than others and are therefore able to generate a better rapport with witnesses. This raises questions about the validity of research into the enhanced cognitive interview. Research into the cognitive interview has obvious ‘practical applications’ This research can help the police to aid the recall of witnesses. In practice however many police forces report that they only use ‘recall everything’ and ‘context reinstatement’ in police interviews as they believe the ‘change perspective’ and ‘reverse order’ techniques confuses witnesses.