Research on coding and evaluation
Alan Baddeley investigated the process of coding (the format in which information is stored in the memory stores) by giving different lists of words to 4 groups of participants to remember:
- Group 1 (acoustically similar): words sounded similar, eg. cat and can
- Group 2 (acoustically dissimilar): words sounded different, eg. pit and cow
- Group 3 (semantically similar): words with similar meanings eg. great and large
- Group 4 (semantically dissimilar): words had different meanings, eg. good and hot
Participants then were shown the original words and were asked to recall them in order. When they had to do so immediately after hearing them, they tended to do worse with acoustically similar words.
If they were to recall the word list after a time interval of 20 minutes. they did worse with the semantically similar words, suggesting that information is coded semantically in LTM.
A criticism of Baddeley's research into coding is that he used artificial stimuli rather than meaningful material. This means caution should be taken when generalising the findings to different memory tasks. For example, when processing more meaningful information, people may use semantic coding even for STM tasks. This suggests the findings from this study have a limited application.
Research on capacity and evaluation
Joseph Jacobs developed a technique to measure digit span in order to research capacity. For example, 4 digits would be given to a participant and then the participant is asked to recall these in order, out loud. If this is correct, 5 digits are read and so on until the participants cannot recall the order correctly. This determines the individual's digit span.
George Miller made observations of everyday practice, noticing many things come in sevens: 7 notes on the musical scale, 7 days in a week, 7 deadly sins and so on. This suggests the capacity of STM is 7+-2. Miller also noted that people can recall 5 numbers as well as 5 letters, by chunking - grouping the sets od numbers or digits in chunks.
A limitation of Jacob's study is that it was conducted a time ago when psychological research lacked control. For example, some participants might have not performed as well due to distractions. This means confounding variables were not controlled. However, other research has been carried out that supports this study's validity.
A limitation of Miller's research is that he may have overestimated the capacity of STM. Cowan concluded that the capacity of STM was around 4 chunks, suggesting that the lower end of Miller's estimate seems more appropriate than the upper end.
Research on duration and evaluation
Margaret and Lloyd Peterson tested duration of STM on 24 undergraduates. Each student took part in 8 trials. On each trial the student was given a consonant syllable like YCG to remember as well as a 3-digit number. The student was then asked to count back from that 3-digit number until told to stop. This was to prevent any mental rehearsal of the consonant syllable. On each trial they were told to stop at a different length of time - 3,6,9,12,15,18 seconds. This was the retention interval. The findings have showed that the % of correct responses decreased with the increase in retention interval.
A limitation of Peterson&Peterson's study is that the stimulus material was artificial. The task did not reflect most real-life memory activites, meaning the study lacked external validity. The study however is somehow relevant as sometimes we try to remember fairly meaningless things such as phone numbers.
Another limitation of the study is that it lacked internal validity. Things are forgotten in STM due to lack of rehearsal, also known as spontaneous decay, or due to displacement - new information would push out whatever was already there. In the study, participants counted down during the retention interval, meaning that the information was displaced by new information.
Research on duration and evaluation 2
The duration of LTM has been investigated by Harry Bahrick and his colleagues. They studied 392 participants from the American state of Ohio aged 17--74. High schoool yearbookswere obtained either from the participants themselves or directly from schools. Recall was tested in either:
- Photo recognition test consisting of 50 photos, some from the participant's yearbook
- Free recall test where participants recalled all the names of their graduating class
The findings showed that participants who were tested within 15 years of their grduation were around 90% correct in the photo recognition. After 48 years, this declines to 70% to photo recognition. Free recall was less good than recognition. After 15 years, this was about 60% accurate, dropping to 30% after 48 years. This shows that the duration of LTM is very long.
A strength of Bahrick et al.'s study is that it has higher external validity. This is because real-life meaningful memories were studied. When studies on LTM have been conducted with meaningless memories, recall rates were lower, eg. Shepard. The downside to this is that confounding variables are not controlled, for example the participants could've looked at their yearbook photos and rehearsed their memory over the years.
Multi-store model of memory (MSM) description
Atkinson and Shiffrin's multi-store model of memory describes how information flows through the memory system.
Firstly, information from the environment (5 senses) enters the sensory register. This store isn't one but several, one for each of our five senses. The two main stores are called iconic memory and echoic memory. Iconic is visual info coded visually, echoic is auditory infromation coded acoustically. The duration is less than half a second, compared to the capacity which is very high. If attention is paid to the information, it passes further into STM.
Short-term memory has a duration of 30 seconds and has a capacity of 7+-2. Some research suggests the capacity is more like 5 rather than 9. Maintenance rehearsal occurs when we rehearse the material over and over. This means we can keep it in STM longer, and if we rehearse it long enough, it passes into LTM.
Long-term memory has an unlimited capacity and an unlimited duration. For example, Bahrick et al. found that participants were able to recognise names and faces of their school classmates almost 50 years after graduating. LTM also tends to be coded semantically. When we want to recall the information, it needs to be transferred back into STM through retrieval. None of our memories are recalled directly from LTM.
Multi-store model of memory (MSM) evaluation
-Research support. Baddeley found that we tend to mix up words that sound similar when using STMs, but we mix up words that have similar meanings when we use our LTMs. The strength of this study is that it shows that coding in STM is acousting and in LTM it is semantic. So the two stores are different, supporting MSM's view that the two memory stores are separate and independent.
-There is more than one type of STM. Shallice and Warrington studied a patient with amnesia known as KF, and found that KF's short-term memory for digits was poor when they read them out loud to him, but his recall was much better when he was able to read them to himself. This shows that there could be another STM store for non-verbal sounds, like noises. The singular STM store is a limitation of MSM as research shows there must be at least one more to process visual information and one to process auditory information.
-There is more than one type of rehearsal. Craik and Watkins found that MSM's assumption of the more you rehearse the more goes into LTM is wrong. What matters about rehearsal isn't the amount, but the type. Maintenance rehearsal does not transfer information into LTM, it just maintains it in STM. Elaborative rehearsal is needed for long-term storage, occuring when you link information to existing knowledge. This is a limitation of MSM as it is another research finding that can't be explained by the model.
Multi-store model of memory (MSM) evaluation 2
-Artificial materials have been used in research studies supporting MSM. In everyday life, we form memories related to useful things like people's faces and their names, but a lot of the research studies supporting MSM used things like digits and letters. Consonant syllables for example, have absolutely no meaning.
-There is more than one type of LTM. Like STM, LTM is not a unitary store. For example, we have one LTM store for facts about the world, and another for our memories of how to ride a bicycle.
Types of long-term memory
Episodic memory is a long-term memory store for personal events. It refers to our ability to recall events from our lives. First of all, they will be time-stamped as you'll remember when they have happened. Secondly, your memory of a single episode will contain several elements, and all of them are interwoven to produce a memory. Then you need to make a conscious effort to recall episodic memories.
Semantic memory is a long-term memory store for our knowledge of the world. It refers to our ability to recall facts in the broadest sense. It includes the knowledge of things like applying to university and the meaning of words. Your semantic memory contains your knowledge of concepts and meanings. These memories are not time stamped. It is less personal and more about the facts we all share.
Procedural memory is a long-term memory store for our knowledge of how to do things. This includes actions and skills and we recall those without conscious awareness or great deal of effort. These are the skills we might find difficult to explain to someone else.
Evaluation of types of long-term memory
Clinical evidence such as studies of HM and Clive Wearing show that episodic memory in both men was severely impaired due to amnesia. They had great difficulty recalling events but their semantic memories remained unaffected. Their procedural memories were also intact. So HM wouldn't recall stroking a dog half an hour earlier but he wouldn't need to have the concept of 'dog' explained to him over and over again. This supports Tulving's view that there are many memory stores in LTM, and one can be damaged with others remaining unaffected.
Neuroimaging evidence - brain scans show that different types of memory are stored in different parts of the brain. Tulving et al. found that episodic and semantic memories were recalled from the prefrontal cortex, with the left prefrontal cortex being involved in recalling semantic memories and episodic memories being recalled from the right prefrontal cortex. This is a strength as it proves that there is a physical difference in the different types of LTM. Further studies have supported this, increasing the validity of this finding.
Real-life applications by Belleville et al, who showed that episodic memories could be improved in older people who had a mild cognitiveimpariment. The trained participants did better on a test of episodic memory rather than a controlgroup. This shows that the benefit of knowing different types of LTM is that it enables for specific treatments to be developed.
Evaluation of types of long-term memory 2
Problems with clinical evidence, such as clinical studies of HM and Clive Wearing are not perfect. There is a serious lack of control over variables and therefore it is difficult to establish whether this is really how memory is meant to work normally.
Researchers argue over the number of LTM stores. Tulving argues that it is three types, whereas Cohen and Squire only accept procedural memory. they argue episodic and semantic memories are stored in a single store, called the declarative memory. Procedural memories are non-declarative.
The working memory model description
The working memory model is only regarding short-term memory, concerned with how it functions. The model consists of 4 main components.
1. Central executive - Co-ordinates the activities of the three subsystems in memory. It also allocates processing resources to those activities. The central executive has a very limited storage capacity.
2. Phonological loop - One of the slave systems, dealing with auditory information and preserves the order in which the information arrives. It consists of the phonological store (storing the words you hear) and the articulatory process (allows for maintenance rehearsal).
3.Visuo-spatial sketchpad - Second slave system, storing visuo-spatial information when required. It is subdivided into the visual cache (stores visual data) and the inner scribe (records arrangment of objects in the visual field).
4. Episodic buffer - Third slave system, used for temporary storage of information, integrating the visual,spatial and verbal information processed by other stores and recording events that are happening. It has a limited capacity of about 4 chunks and it links working memory to LTM and wider cognition processes such as perception.
The working memory model evaluation
- Clinical evidence coming from Shallice and Warrington's study of KF who suffered brain damage. KF had poor STM ability for verbal information but he processed visual information normally. This suggests that it was only his phonological loop that was damaged, supporting the existence of separate visual and acoustic stores.On the other hand, this is unreliable as the case was unique.
- Studies regarding dual-task performance support the separate existence of the visuo-spatial sketchpad. Baddeley showed participants had more difficulty carrying out 2 visual tasks than doing both a visual and verbal task at the same time. This means that there must be a separate slave system (VSS) that processes visual input.
- There is a lack of clarity over the central executive. Baddeley himself recognised it, saying that the central executive is the most important component but also one that is the least understood. This suggests that the WWM has not been fully explained as of yet.
- Baddeley demonstrated that people find it more difficult to remember a list of long words compared to short words (world length effect). This is due to the limited space for rehearsal in the articulatory process. The word length effect disappears when a person is given an articulatory suppression task - a repetitive task tying up the articulatory process. This supports the effect of the phonological loop.
The working memory model evaluation 2
- Brain scanning studies have supported the WMM. Braver et al. gave his participants tasks that involved the central executive while having a brain scan. Greater activity was spotted in the left prefrontal cortex, and the activity increased as the task became more difficult.
Explanations for forgetting - interference
Interference is the process of forgetting due to one memory blocking another, causing one or both to be distorted or forgotte. This has been mainly proposed as an explanation for LTM rather than STM because interference makes it harder for us to locate the memory in the LTM store.
There are two types of interference:
1. Proactive interference which occurs when an older memory interfers a newer one. For example, a teacher struggling to recall the names of her current class.
2.Retroactive interference which occurs when a newer memory interfers with an older one. For example, the teacher learnt so many names this year, she forgot the ones from last year.
Explanations for forgetting - interference 2
The interference seems to be worse when the memories are similar, which has been discovered and researched by McGeoch and McDonald.
They studied retroactive interference by changing the amount of similarity between two sets of materials. Participants were asked to learn the words until they remembered them with 100% accuracy, and then they learned a new list. There were six groups:
1. Synonyms - words with same meanings
2. Antonyms - words with opposite meanings to originals
3. Words unrelated to original ones
4. Nonsense syllables
5. 3 digit numbers
6. No new list - just rest.
The findings showed that when the participants recalled the original list of words, their performance depended on the nature of the second list. Most similar material had the worst recall, showing that interference is strongest with memories that are similar.
Evaluation of interference
- Evidence from lab studies shows that both types of interference are likely to be the common ways we forget information from LTM. This is a positive as lab experiments control the effects of irrelevant influences, giving us confidence that interference is a valid explanation for forgetting.
- Artificial material was used, meaning that it does not apply to real-life situations. Learning lists of actual words is more realistic than learning lists of consonant syllables, but this is still far from real-life. Interference therefore might not be as likely an explanation for forgetting in everyday life as in the lab.
- Real-life studies such as that by Baddeley and Hitch supports the application of interference to real-life situations. They asked rugby players to remember the names of the teams they played in a season, week by week. The results have showed that accurate recall did not depend on how long ago the matches took place but the actual number of games played in the meantime. A player's recall of a team from three weeks ago was better if no games were played since then.
- Lab experiments are designed to maximise the possibility of interference. For example, the time periods between learning and recall are short, so a participant might have to learn one list of words, then one 20 minutes later, and recall one of them a few minutes after that.
Evaluation of interference 2
- Interference effects may be overcome using cues. Tulving and Psotka gave participants 5 lists of 24 words, each organised into six categories. The categories were not explicit, but it was presumed they would be obvious to participants. Recall was 70% for the first word list but this fell as participants were given each additional list to learn. At the end they were given a cued recall test, in which they were told the names of the categories as a clue. In this condition, recall rose to 70% yet again.
Explanations for forgetting - retrieval failure
Retrieval failure is a form of forgetting that occurs when we don't have necessary cues to access memory
Endel Tulving investigated one cause behind retrieval failure - encoding specificity principle (ESP). This states that if a cue helps us to recall information, it has to be present at encoding and retrieval. Some cues are linked to the material that is to be remembered in a meaningful way, for example when we hear STM, we recall all sorts of information about it.
There are two types of forgetting that are linked to meaningless cues:
Context-dependent forgetting has been investigated by Godden and Baddeley. In this study, divers learnt a list of words, underwater or on land. Then, they were asked to recall these either underwater or on land, thus creating four conditions: learn land recall land, learn land recall underwater, learn underwater recall land, learn underwater recall underwater. The findings showed that accurate recall was 40% lower in the non-matching conditions, suggesting external cues available at learning were different to those at retrieval, leading to retrieval failure.
Explanations for forgetting - retrieval failure 2
State-dependent forgetting has been investigated by Carter and Cassaday. They gave anti-histamine drugs to their participants - those had mild sedative effects, making the participants slightly drowsy, a different state from the normal one of being awake and alert. The participants were made to learn lists of words and passages of prose, and thus recall the information in four conditions: learning on drug recall when on drug, learning on drug recall when not on it, learning not on drug recall when on drug and learn not on drug recall not on drug. The findings have showed that where the conditions were mismatched, the performance was much worse, suggesting that where the cues are absent, there is more forgetting.
Evaluation of retrieval failure theory
- There is supporting evidence for the existence of the retrieval failure explanation. The studies carried out are just examples of this research. Eynseck goes as far to argue that retrieval failure is the main reason for forgetting from LTM. This is a strength as validity of the explanation is increased.
- Context effects have been questioned by Baddeley, who argues that they aren't very strong, especially in real-life. To really study the effects of different contexts, they would really have to be different. Learning something in one room and recalling it in another is unlikely to show a lot of forgetting. This shows that the theory cannot be applied to real-life.
- Another limitation of the context effects is that cues only affect memory when you test it in a certain way. For example, Godden and Baddeley replicated their underwater experiment but with a recognition test instead. There was no context-dependent effect, as the performance was same in all four conditions.
- The encoding specificity principle (ESP) cannot be measured. We assume that a cue produces the successful recall of a word because it has been encoded at the time, but that's just an assumption, so a conclusion cannot be established.
- There are certain useful real-life applications e.g. most of us having the following experience: we are upstairs in our bedroom and think we need to get so and so from downstairs, only to go downstairs and forget what we came for, but when we go back upstairs, we remember it again.
Factors affecting eyewitness testimony: Misleading
Eyewitness testimony is referred to as the ability of people to remember the details of events such as crimes. The accuracy of EWT is affected by misleading information, leading questions and anxiety.
Misleading information is incorrect information given to the eyewitness after the event.
Leading questions have been investigated by Loftus and Palmer. In their research, they asked students to watch film clips of car accidents. They were to answer questions regarding the accident, with the critical question being ' About how fast were the cars going when they hit each other?'. The verb in this sentence was changed for every group of participants, and included contactd, bumped, collided, smashed and hit. The mean score of the participant group with each verb was estimated, and the results showed that the verb contacted had the lowest estimate of 31.8, whereas smashed had the mean of 40.5.
The response bias explanation suggests that the wording of the question has no real effect on the participants' memories, but influences the way they answer. Loftus and Palmer carried out a second experiment to support the substitution explanation. Participants who heard the word smashed were more likely to report seeing broken glass when there was none, than those who heard hit.
Factors affecting eyewitness testimony: Misleading
The second factor affecting the accuracy of EWT is post-event discussion. This occurs when there is more than one witness to an event, so they might discuss what they have seen and influence the accuracy of each other's recall of the event.
To support this, Fiona Gabbert and her colleagues studied participants in pairs. Each participant watched a video of the same crime but filmed from different points of view, so each participant could see elements in the even that the other could not. For example, only one could see the title of the book carried by the woman. Then, both participants cohesively discussed what they had seen before being assessed individually.
The results show that 71% of the participants mistakenly recalled the events that weren't seen by them in the video, but the ones that they picked up during the discussion. In the control group, where there was no discussion, the score was 0%. This concludes that witnesses often go along with one another, referring to this as memory conformity.
Misleading information evaluation
- It has useful real life applications. For example, Loftus believes that the leading questions can have a distorting effect on memory strong enough for the police officers to be forced to phrase their questions to eyewitnesses carefully. This means that research into EWT makes a positive difference to the lives of real people by improving the legal system.
- A negative is that the research uses artificial tasks. Loftus and Palmer had their participants watch clips of car accidents which is a very different experience to that of witnessing a real one. The emotions aren't quite the same, meaning that these studies give us very little about how leading questions can affect EWT.
- There is evidence for individual differences, as older people are less accurate than younger people, proven by Anastasi and Rhodes. However, all age groups were more accurate when identifying of their own age group (own age bias).
- Demand characteristics also play a part - the answers given in lab studies of EWT might've been false. So if you've seen a film of street robbery and you were asked the question ' did you see the blue car ' even though there wasn't one, you'd answer yes as that seems more useful.
- In studies, the consequences of giving false testimonies are more forgiving compared to real life.
Factors affecting eyewitness testimony: Anxiety
Anxiety is a state of emotional and physical arousal. Emotions can include having worried thoughts whereas physical changes can include an increased heart rate.
Anxiety can have a negative effect on recall:
Anxiety creates physiological arousal in the body which stops us from paying attention to important cues, making recall worse. Studying anxiety and EWT was approached by looking at the effect of weapons, by Johnson and Scott. They led participants to believe they were taking part in a lab study. While seated, they heard an argument in the next room. In the 'low anxiety' condition, a man walked through carrying a pen with grease on his hands. In the 'high anxiety' condition, other participants heard the same argument but this time accompanied with the sound of breaking glass. A man walked out of the room holding a paper knife covered in blood.
The findings showed that 49% of the participants who saw the man with the pen were able to identify him from a set of 50 photos, and the figure for those who saw the man holding a blood-covered knife was just 33%. This tunnel theory argues that witness's attention narrows to focus on a weapon because it is a source of anxiety.
Factors affecting eyewitness testimony: Anxiety 2
However, anxiety can also have a positive effect on recall during EWT:
The stress of witnessing a crime creates arousal within the body. The fight-or-flight response is triggered and as a result, the alertness improves our memory for the event as we are aware of the cues in the situation.
This approach was tested by Yuille and Cutshall, who conducted a study of a real-life shooting in a gunshop in Canada. The shop owner shot the thief dead. There were 21 witnesses (13 agreed to take part in the study). The interviews were held 4-5 months after the incident and those were compared with the original police interviews made at the time of the shooting. Accuracy was marked by the number of details reported. The witnesses were asked to rate their stress levels at the time of the incident on a scale of 1-7, as well as being asked whether they had any emotional problems since.
The witnesses were accurate in their accounts and there was little change in the accuracy over 5 months. Some details were less accurate such as the recollection of colour of items and age. The participants who reported the highest levels of stress were the most accurate.
Factors affecting eyewitness testimony: Anxiety 3
According to Yerkes and Dodson, the relationship between emotional arousal and performance looks like an inverted U.
Kenneth Deffenbacher applied the Yerkes-Dodson Law to EWT. Low levels of anxiety produce lower levels of recall accuracy, but memory becomes more accurate as the level of anxiety increases. However, the maximum point of accuracy is reached where there is just enough anxiety - if an eyewitness experiences any more stress than this, then their recall declines massively.
- The study by Johnson and Scott regarding weapons might not be relevant. It might test surprise rather than anxiety. Pickel conducted an experiment using scissors, a handgun, a wallet or a raw chicken as the hand-held items in a hairdresing salon video. EW accuracy was significantly poorer in the high usualeness conditions like the chicken and the handgun.
- Field studies sometimes lack control - researchers usually interview real life EW after the event, and all sorts might have happened to the participants in the meantime, for example discussions with other people about the event. This is a limiation as extraneous variables might be responsible for the accuracy.
- There are ethical issues in research into anxiety - it subjects people to psychological harm purely for the purpose of the research. Psychologists interview people who have already witnessed the event, so there is no need to create it. It doesn't challenge the findings from studies but questions the need for it.
- The inverted U explanation is too simplistic - anxiety has many elements; cognitive, behavioural, emotional and physical. It assumes that poor performance is only linked to physiological arousal.
- Demand characteristics operate in lab studies - the participants will be aware of watching a filmed crime for a reason. Chances are, they'll work it out for themselves that they are going to be asked questions about what they saw.