Memory

Coding = the format in which information is stored in the various memory stores 

Baddeley (1966) Coding in STM and LTM

Procedure

• group 1: accoustically similar words (eg. cat, cab)
• group 2: dissimilar (eg. pit, few)
• group 3: semantically similar (eg. large, big)
• group 4: dissimilar (eg. good, hot)

Findings

• Immediate recall worse with accoustically similar words = STM is acoustic
• Recall after 20 mins worse with semantically similar words = LTM is semantic

limitation - words in the study had no personal meaning to the ppts. When processing more meaningful info, people may use semantic coding even for STM tasks - limited application

Capacity = the amount of memory that can be held in a memory store

Jacobs (1988) Capacity of STM

• Digit span: Researcher reads 4 digits and ppts recalls. no. of digits keeps increasing until ppt can no longer recall the order correctly
• On average, ppts could repeat back 9.3 numbers and 7.3 letters immediately after

limitation - conducted a long time ago when studies lacked control of extraneous variables - invalid results. However, othere research supports the findings, supporting its validity.

Miller (1956) Capacity of STM

• made observations of everday practice, eg. noted that things come in sevens: 7 notes on music scale, 7 days of the week, 7 deadly sins etc.]
• The span of STM is approx 7 items (plus or minus 2) but can be improved by chunking - grouping sets of digits/letters into meaningful units

limitation - Cowan et al (2001) concluded that the capacity of STM was approx 4 chunks - suggests the lower end of Miller's estimate (5 items) is more appropriate than 7

Duration = length of time info can be held in the memory

• Peterson and Peterson (1959) Duration of STM
• 24 students, each took part in 8 trials 
• given constant syllable and a 3 digit number - told to count backwards for 3, 6, 9, 12, 15, or 18 secs (different on each trial) - avoids any mental rehearsal of constant syllable
• 3 sec interval = recall 80% correct, after 18 secs = recall fell to 3%
• suggests duration of STM without rehearsal is about 18 to 30 secs.
• Limitation - memorising constant syllables doesn't reflect real life memory activities - lacks external validity. However, we try to remember phone no. - not totally irrelevant

• Bahrick et al (1975) Duration of LTM
• 392 Americans aged between 17 and 74
• recognition test: 50 photos from ppts high school year book
• free recall test: ppts listed names of their graduating class
• photo recognition - tested 48 yrs after grad = 70% accuarcy - free recall less accuarte
• strength - meaningful memories studied - when lab studies done with meaningless pics - recall lower. However, confounding variables not controlled - looked at year book photos and rehearsed memories over the years

• Sensory Register (SR) - stimulus from environment passes thru SR along with other sights, sounds etc. SR has 5 parts, each for the different senses
• duration: very brief - less than half a sec
• capacity: high, eg. over a hundred million cells in the eye, each storing data
• coding: depends on the sense

transfer from SR to STM - only info paid attention to passes onto STM

• STM 
• duration: approx 18 to 30 secs, unless info is rehearsed
• capacity: between 5 and 9 items 
• coding: acoustic

transfer from STM to LTM - maintenance rehearsal occurs when we rehearse material to ourselves. Rehearsed long enough = passed on to LTM

• LTM - recall material from LTM - has to be transferred back to STM thru process retrieval
• duration: potentially up to a life time
• capacity: potentially unlimited
• coding: semantic

Baddeley (1966) found we tend to mix up words that sound similar (acoustics) when using our STM, but mix up words with similar meanings (semantics) when we use LTM - clear difference between STM and LTM - supports MSM view that the 2 stores are separate and independent

Shallice and Warrington (1970) studied KF, a patient with amnesia - his STM for digits was poor when they read them out loud, but his recall was better when he read the digits himself - MSM states there is one STM, but KF's study suggests there must be more than one STM store to process visual info and another for auditory info - WMM is better explanation bcos includes separate stores

Craik and Watkins (1973) argued there are 2 types of rehearsal: maintenance and elaborative - MSM only includes maintenance - elaborative is need for LTM storage - occurs when you link info to your existing knowledge - serious limitation; another research finding that cant be explained by the model

Researchers often asked ppts to recall digits, letters etc. - in everyday life we form memories related to useful things eg. peoples faces, names etc - MSM lacks external validity; doesn't reflect how memory works in everyday life

LTM Store 1: Episodic Memory

• stores events (episodes) from our lives; daily happenings, eg. recent trip to the dentist
• time-stamped
• involve several elements, eg. people, places, objects and behaviours
• have to make a conscious effort to recall them

LTM Store 2: Semantic Memory

• stores our knowledge of the world, eg. the taste of an orange
• not time-stamped, eg. we don't remember the first time we tasted an orange
• less personal; knowledge we all share

LTM Store 3: Procedural Memory

• stores memories for actions and skills, eg. driving a car
• recall occurs without conscious awareness - find hard to explain to someone else

Clinical studies of amnesia, such as HM and Clive Wearing, showed both difficulty recalling events that had happened to them in the past, but their semantic memories were relatively unaffected eg. HM cant remember stroking a dog half an hour later, but does know what a 'dog' is - supports the view that there are different LTM stores bcos 1 store can be damaged leaving others unaffected

Tulving et al (1994) had ppts perform various memory tasks while their brains were scanner with a PET scanner - episodic and semantic memories located in prefrontal cortex; semantic left side, episodic right side - supports validity that there are different types of LTM

Belleville et al (2006) found that episodic memories can be improved in older people with mild cognitive impairments. Training led to improvements (compared to control group) - highlights benefit of distinguishing between different types of LTM - it allows specific treatments to be developed

Clinical evidence, such as HM and Clive Wearing, lack control of variables, eg. cannot control the precise location of the brain damage or personality variables - difficult to generalise findings to determine exact nature of LTM

WMM is a model of how the STM is organised and how it works. 

• Central executive allocates slave systems
• attentional process that monitors incoming data and allocates slave systems to tasks
• very limited storage capacity

• Phonological loop deals with auditory info + preserves it in the order it arrives
• Phonological store: stores the words you hear
• Articulatory process: allows maintenance rehearsal to keep them in WM while needed

• Visuo-spatial sketchpad stores visual and/or spatial info when required
• Logie (1995) subdivided the VSS into:
• Visual cache: stores visual data
• Inner scribe: records arrangement of objects in visual field

• Episodic buffer is a tempory storage for info
• integrates visual, spatial and verbal info from other stores
• maintains sense of time sequencing 
• links to LTM

Shallice and Warrington (1970) carried out a case study of patient KF who had brain damage - found to have poor STM ability for verbal info but could process visual info - phonological loop has been damaged and others stores are intact - supports there are separate visual and acoustic stores - However, evidence from brain damaged patients may be unreliable bcos concerns unique cases of patients who have had traumatic experiences

Baddeley et al (1975) found ppts had more difficulty doing 2 visual tasks (eg. tracking a light and describing the letter F) than doings a visual and verbal task at the same time - greater difficulty is both visual tasks compete for the same resources - doing a visual and verbal simultaneously = no competition - provides evidence for visuo-spatial sketchpad - better explanation than MSM

Cog psychologists suggest the CE is unsatisfactor / doesn't explain anything - CE should be more clearly specified; some psychologists believe should consist of separate components - WMM hasn't been fully explained / is over-simplified

Baddeley et al ppts had more difficulty remembering long words than short (word length effect) - bcos limited space for rehearsal in articulatory process (2 secs) - word length effect disappears if ppts given repetitive task - proves how the articulatory process works

• forgetting occurs in the LTM bcos can't access memories even thought they are available
• interference = when two pieces of info are in conflict 

• Proactive interference = old memory interferes with new memory
• Retroactive interference = new memory interferes with old memory

Interference is worse when memories are similar - may be bcos:

• in PI previously stored info makes new info more difficult to store
• In RI new info overwrites previous memories which are similar

Key Study: McGeoch and McDonald (1931) effects of similarities

Procedure

• ppts asked to learn list of words to 100% accuracy
• then given a new list to learn - new words varied in degree to which it was similar to other list:
• group: synonyms 
• group 2: antonyms
• group 3: unrelated
• group 4: non-sense syllables
• group 5: 3 digits numbers
• group 6: no new list

Findings

• performance depended on nature of second list - synonyms produced worst recall
• when ppts given very different list, eg. 3 digit numbers - mean no. of items recalled increased
• shows interference is stronger when memories are similar

many lab experiments have been carried out into interference, eg. McGeoch and McDonald - most studies show both types of inteference are likely causes of forgetting from the LTM - lab experiments control effects of extraneous variables = interference is valid explanation

material used is word lists - more realistic than consonant syllables, but still different from everyday life, eg. peoples faces, birthdays etc. - use of artificial materials interference more likely in the lab - cannot be generalised to everyday forgetting

Baddeley and Hitch (1977) asked rugby players name teams that they played that season week by week - accurate recall didnt depend on how long ago the match took place, more important was no. of games played in the mean time - shows interference can apply to some everyday situations

time periods between learning lists of words and recall are quite short in lab studies - research reduces whole experience of learning into short time period = doesnt reflect how we learn / remember most info in life - role of interference may be exaggerated - cannot be generalised

when info is initially placed in memory, associated cues are also stored - cues are not available at time of recall = might not be able to access memories 

Tulving (1983) suggested that cues help retrieval if same cues are present at encoding and at retrieval - Encoding specificity principle (ESP) - the closer the retrievel cue to origional cue, the better the cue works

some cues are linked to the material to be remembered in a meaningful way, eg. the cue STM may lead you to recall info about short term memory

other cues are also encoded at time of learning but not in meaningful way

• context dependent forgetting: when memory of retrieval is dependent on external / environmental cue, eg. weather / place
• state-dependent forgetting: when memory retrieval is dependent on internal cue, state of mind, eg. upset, being drunk

Procedure

• cues were the contexts where learning and recall took place - on land or under water
• deep sea divers learned word lists and were later asked to recall
• group 1: learn on land, recall on land
• group 2: learn on land, recall underwater
• group 3: learn under water, recall on land
• group 4: learn underwater, recall underwater

Findings

• When environmental context of learning and recall did not match, ie. conditions 2 and 3, accurate was 40% lower than when did match, ie. 1 and 4
• when external cues at learning were different to ones at recall = retrieval failure bcos lack of cues
• study demonstrates context dependent forgetting bcos info not accessible, ie. was forgotten, when context at learning did not match recall

There is an impressive range of evidence to support retrieval failure; alongside Godden and Baddeley's research, Eysenck (2010) goes so far as to argue that retrieval failure is perhaps the main reason for forgetting in LTM - studies conducted in real-life situations as well as highly controlled lab conditions = increases validity

Baddeley argued that different contexts have to be very different before an effect is seen, eg. on land vs in water - learning in one room and recalling in another is unlikely to result in forgetting - environments aren't different enough - real life applications of RF don't fully explain forgetting

Godden and Baddeley (1980) replicated their underwater experiment using a recognition test instead of recall - no context dependent effect - performance was the same in all 4 conditions - limits RF as an explanation for forgetting  - presence/absence of cues only affects memory when u test recall rather than recognition

Are some real life applications, eg. going up stairs and forgetting what you went up for, so you return back downstairs to remember - basic principle of the cognitive interview, a methof of getting eyewitnesses to recall more info about crimes by using a technique called 'context reinstatement'       

• response bias: wording of a question has no enduring effect on an eyewitness's memory of an event, but influences the kind of answer given
• substitution explanation: wording of a question does effect eyewitnesses memory; it interferes with its origional memory, distorting its accuracy

Key study: Loftus and Palmer (1974) Leading questions

• 45 students watched film clips of car accidents and then answered questions about speed
• Critical question: 'About how fast were the cars going when they hit each other?'
• 5 groups, each given different verb: hit, contacted, bumped collided or smashed

Findings

• 'contacted' = mean estimated speed of 31.8mph, 'smashed' = mean of 40.5mph
• leading question biased eyewitnesses recall of an event

limitation of study: ppts watched film clips, very different than witnessing a real accident - Yuille and Cutshall found witnesses of a real traumatic armed robbery had very accurate recall after 4 months - shows findings from artificial tasks cannot be generalised to real crimes

• memory contamination: when co-witnesses discuss a crime, individuals mix (mis)information from other witnesses with their own memories
• memory conformity: witnesses go along with eachother to win social approval or because they believe the other witnesses are right

Key study: Gabbert et al (2003) Post-event discussion

• paired ppts watched a vid of the same crime, but filmed so each ppt could see elements in the event that the other could not
• paired ppts discussed what they had seen on the vid before completing a test of recall

Findings 

• 71% ppts recalled aspects of the event they did not see, but had picked up in the PED
• in a control group (no discussion) there were no errors in what they saw

The research has led to important practical uses for police officers and investigators - consequences of inaccurate EWT can be very serious - Loftus claimed leading questions can have a distorting influence on memory - police officers need to be careful about how they phrase questions when interviewing eyewitnesses - research led to improving the legal system

Anastasi and Rhodes found older people were less accurate than younger people when giving eyewitness reports - however also found all age groups were more accurate when identifying people of their own age (age-bias) - research studies often use younger people as the target to identify - some age groups seem less accurate, but not really the case

ppts will want to be helpful therefore may guess the answer to questions they don't know - 'did you see the blue car?' may answer yes, even tho there was no blue car, as seems more helpful answer - may suffer from demand characteristics = challenges validity

Foster et al argue that what you remember as an eyewitness can have important consequences in the real world, but not in research studies - real eyewitnesses search their memory more thoroughly bcos testimony may lead to a successful conviction - EWT accuracy may be greater in the real world 

• ppts sit in waiting room believing they were going to take part in a lab study
• each ppt heard an argument in the next room, 2 conditions:
• low anxiety: a man walked through waiting room carrying a pen with grease on his hands
• high anxiety condition: heated argument accompanied by sound of breaking glass - man walked through waiting room holding paper knife covered in blood
• ppts then asked to identify the man from set of 50 photographs

• Findings 
• low anxiety = 49% ppts able to identify, high anxiety = 33% able to identify
• tunnel theory of memory argues that a witness's attention is on the weapon (weapon focus) bcos it is a source of danger and anxiety

• limitation: ppts may focus on a weapon bcos they are surprised rather than scared - Pickel used scissors, hand gun, wallet and raw chicken as hand-held items in a hair salon, EWT accuracy = poor for high unusualness (gun + chicken) - doesn't explain effects of anxiety
• limitation: creating anxiety in ppts is unethical - psychological harm
• limitation: controlled lab studies suffer from demand characteristics - may guess they are being observed - reduces validity

• real life crime: gun-shop owner shoots thief dead
• 21 witnesses, 13 agreed to participate and were interviewed 4-5 months after incident
• accounts compared to police interviews at the time of shooting
• witnesses how stressed they felt at time of incident

Findings

• witnesses were very accurate and there was little change after 5 months - some details less accurate eg. colours of items, age/weight/height
• ppts who reported the heighest levels of stress were most accurate with 88%, compared to 75% for the less stressed group

limitation: real life witnesses interviewed sometime after the event - in this time eyewitnesses discuss the event with others, read or view accounts in the media etc - extraneous variables may be responsible for the (in)accuracy of recall, not anxiety - difficult to isolate these variables

benefit: real life studies beneficial - no need to create anxiety

'Inverted U theory': Yerkes and Dodson (1908) argue that the relationship between performance and arousal/stress is curvlinear rather than linear

affects memory: Deffenbacher (1983) found that lower levels of anxiety sis produce lower levels of recall accuracy. Recall accuracy increases with anxiety up to an optimal point. A dratic decline in accuracy is seen when an eyewitness experiences more anxiety than the optimal point

limitation: inverted U theory is limited because it is too simplistic - anxiety is difficult to define and measure bcos it has many elements - cognitive, behavioural, emotional and physical - the inverted U theory assumes that one of these (physical) is linked to poor performance - the explanation fails to account for other factors

• Fisher and Geiselman (1992) claim that EWT could be improved if the police use techniques based on psychological insights into how memory works
• they called it the cognitive interview to indicate its foundation in cognitive psychology
• Rapport (understanding) is established with the interviewee

• 1. Report everything: witnesses are encourage to report every detail, even if it seems irrelevant or the witness is not confident about it
• trivial details may be important or may trigger other memories 

• 2. Reinstate the context: witness returns to orgional crime scene 'in their mind' and imagines the environment and their emotions - based on context-dependent forgetting

• 3. Reverse the order: events recalled in different chronological order eg. end to beginning, to prevents using expectations of how events must have happened and to prevent dishonesty

• 4. Change perspective: witnesses recall event from others perspective - prevents influence of expectations and schema on recall

Milne and Bull found that each individual element of the CI was equally valuable - however, they found that a combination of 'report everything' and 'context reinstatement' produced better recall than any of the techniques individually - these two elements at least should be used to improve police interviewing even if the full CI isn't used

Police reluctant to use CI bcos it's more time-consuming - more time is needed to establish rapport with the witnesses to allow them to relax - Kebbel and Wagstaff point out that the CI also requires special training, and forces have not been able to provide enough hours - unlikely that the full version of CI is actually used - may explain why police haven't been impressed with it

Studies of the effectiveness of CI inevitably use slightly different techniques - different researchers use variations and police forces evolve their own methods - means it is difficult to draw conclusions about the CI in general

Fisher et al (1987) developed additional elements of the CI:

• focus on the social dynamics of the interactions eg. knowing when to establish eye contact
• ideas such as reducing eyewitnesses anxiety, minimising distractions, getting witness to speak slowly and asking open-ended questions

benefit: A meta-analysis by Kohnken et al combined data from 50 studies - enhanced CI consistently provided more correct info than the standard interview used by police - studies like these indicate that there are real benefits to the police of using the enhanced version of the CI