memory case studies

baddeley (1966)

- gave four groups of participants four different groups of words to recal in order

- acoustically sim/dissim, semantically sim/dissim

- when recalling from stm acoustic was worse, when recaling from ltm after 20 mins semantic was worse

- stm is mostly coded acoustically while ltm is mostly coded semantically

baddeley eval

- showed that we have different memory stores and contributed to the multi store model

- artificial stimuli was used that had no meaning to the participants, we may code semantically in stm for meaningful info

jacobs (1887)

- read out a string of numbers to participant who had to recall them out loud in order

- increase from four until the participant couldn't recall all numbers

- average 7.3

jacobs eval

- has been replicated often and successfully as the study is old and probably didnt have good controls

miller (1956)

- stm capacity 7+-2

- we chunk information to remember it

miller eval

- potentially overestimated, capacity might be 4+-1

lloyd and peterson (1959)

- participants given a consonant syllable to remember and a three digit number to count down from (prevents mental rehearsal)

- participants told to stop counting after 3, 6, 9, 12, 15 or 18 seconds and recall the letters

- 3 seconds = 80% recall, 18 seconds = 3% recall

lloyd and peterson eval

- artificial stimulus used which reduces external validity. not completely meaningless bc we still learn some meaningless things like phone numbers, but still very artificial stimuli

bahrick (1975)

- 392 american participants between 17 and 74 asked to recall the names of people from their graduating class, either photo recognition or free recall

- within 15 years of graduation = 90% accurate photo recall, 60% free

- recall declined with age but still happened

bahrick eval

- hit external validity as it is meaningful stimuli

- tom studies with meaningless pictures had lower recall rates

mcgeoch and mcdonald (1931)

- interference increases with similar memories

- six groups of participants had to learn and recall a different list of words

- synonyms, antonyms, unrelated words, consonant syllables, three digit numbers or no words (control)

- worst recall was the synonyms group

mcgeoch and mcdonald eval

- artificial stimuli means there is poor external validity

- lab based study which means it was highly controlled

• supported by baddeley and hitch = asked rugy players to recall names of teams they played against during a season. players who played in the most recent matches had poorest recall due to interference as all the memories are so similar.

• however tulving = gave participants lists of words organised into categories one at a time and recall became progessively worse bc proactive interference. however, when given a cued recall test the recall rose back up again, so interference may not be permanent memory loss.

godden and baddeley (1975)

CONTEXT DEPENDENT FORGETTING

four conditions for learning and recalling words:

- learn on land, recall on land

- learn on land, recall in water

- learn in water, recall in water

- learn in water, recall on land

- recall 40% lower in non matching conditions

carter and cassaday (1998)

STATE DEPENDENT FORGETTING

four conditions:

- learn on drug, recall on drug

- learning on drug, recall sober

- learn sober, recall sober

- learn sober, recall on drug

- recall worse in mismatched conditions

dependent forgetting eval

- context effects change depending on what type of memory is being tested (recognition, recall etc). when the underwater study was replicated with recognition rather than recall, context effects were not present.

- context effects may not be that present in real life unless the contexts are as severely different as underwater vs on land.

- we cannot really prove that a cue lead to encoding, just assume it did.

• context cues help w everyday recall. we can go to a room and forget why we went there and leave, but when we go back to the room a second time we remember what we needed.

loftus and palmer (1974)

- 45 students watched clips of car accidents and asked questions about it

- 'how fast were the cars going when they hit/contacted/bumped/collided/smashed with each other?'

- 'smashed' had a higher mean speed estimate than 'contacted'

- leading questions can modify the memory of an incident

- substitution explanation = people who reported higher speeds were also more likely to falsely believe there was broken glass

loftus and palmer evaluation

- the study has practical use in the real world, as leading questions can lead to incorrect ewt and modified memories

- in the real world the ewt will have real consequences and so may end up being better than in a lab study where there is no consequences for faulty ewt

- we might be more likely to confidently remember central events than peripheral events, which makes peripheral events less resistant to leading questions

- post even discussion might have more impact on ewt than leading questions

johnson and scott (1976)

- participants believed they were taking part in a lab study and were left in a waiting room

- one group = heard conversation and saw someone leave with a pen and grease on their hands

- other group = heard an argument and glass breaking then someone leave with a knife and blood on their hands

- less people in group two could identify the man from a lineup likely due to tunnel vision on the weapon

johnson and scott eval

• limitation - tunnel vision may be due to surprise rather than anxiety = study in a hair salon where people saw someone with scissors, a handgun or a chicken. worst recall was in the most surprising contexts (handgun and chicken which you wouldn't expect in a salon)

• supporting - a study in London Dungeon found that those in the high anxiety group were less capable of identifying an actor than those in the low anxiety group.

yuille and cutshall (1986)

- natural study of 13 witnesses to a real shooting where a shop owner shot a thief dead

- they rated their stress at the time + were interviewed five months after the incident and these were compared to the police interviews

- recall was fairly accurate ben after five months, with the more stressed people having better recall on average then the less stressed people (88% vs 75%)

yuille and cutshall eval

- witnesses to a real bank robbery in sweden also had more accurate recall the more anxiety they had about the event. direct victims like the bank tellers were the most accurate.

- the gap between the event and the study (4 - 15 months) left time for confounding variables like ped.

atkinson and shiffrin (1968)

• multi store model of memory w three stores

• sensory register w modality specific coding, very high capacity but duration is less than half a second

• stm coded acoustically, lasts about 18 seconds for 7+-2 items, can be prolonged through maintenance rehearsal

• ltm coded semantically, info passes here if maintenance rehearsal done enough, potentially unlimited capacity and duration

atkinson and shiffrin eval

• supported by studies like baddeley study that found worse performance on acousticaly similar words when testing stm

• patient HM supports the msmm, as he had poor ltm after his surgery but his immediate stm recall was fine

• however there is evidence of multiple stm stores, like KF who had poor recall when digits were read aloud to him but better recall when he read them himself

• could be that actually you need to do elaborative rehearsal to transfer info to ltm, not lots of prolonged

clive wearing

• damaged episodic memories but intact procedural memories, knew how to play piano still but when he sees his wife he reacts as though he hasnt seen her in years even though she may have just stepped out the room

• supports multiple ltm stores, but unique case so cant generalise to everyone really

tulving (1994)

• linked semantic and episodic memory to different side of the prefrontal cortex

• conflicting research from other people over what side of the prefrontal cortex has semantic or episodic

baddeley and hitch (1974)

• working memory model

• central executive allocates attention to the two slave systems, doesn’t store information itself

• phonological loop - phonological store = stores words you hear. articulatory process = allows for maintenance rehearsal to keep auditory in loop through repetition. loop = 2 seconds.

• visuo spatial sketchpad - visual cache = stores visual data. inncer scribe = record 3D arrangement of objects.

• episodic buffer = integrates info from the slave systems together, links working memory to ltm

baddeley and hitch eval

• supported by baddeley dual task study = people struggled to do two tasks when both tasks used the same slave system, bc the tasks were competing for the allocated attention.

• central executive is unclear and baddeley himself admits this. could consist of separate components itself, and its unclear how it passes info to ltm.

• supported by patient KF whos vss was intact but pl damaged = had better stm recall when he read digits himself vs when they were read aloud to him.

• however KF had a motorcycle accident which we cant control variables of, so he may have developed other cognitive impairments that we can’t generalise to everybody.

gabbert (2003)

• participants were paired and they each watched a vid of same crime but they saw a different angle to the other person.

• after discussed w their partner and took a recall test, found 71% of participants recalled info they didnt see themself but picked up in the post even discussion

• memory contamination = when other witnesses memories distort our own

• memory conformity = going along with our witnesses even though privately your memory doesnt change

gabbert eval

• demand characters are an example of inacurate ewt due to misinfo, eg a participant may act differently than usual bc they think they know the intentions of the study and want to please the researcher

• study where participants saw a mugging video where the muggers hair was either dark brown or light brown. the participants were allowed ped, and when interviewed after the most common hair colour said was medium brown rather than light or dark