Coding in LTM and STM
BADDELEY
Procedure: Different lists of words to 4 groups to recall in correct order.
- Acoustically similar
- Acoustically dissimilar
- Semantically similar
- Semantically dissimilar
Findings:
- When recalled immediately (STM) - worst with acoustically similar words = acoustically coded.
- When recalled after 20 mins (LTM) - worst with semantically similar words = semantically coded.
Capacity of STM
JACOBS
Procedure: Measured digit span by giving ppts increasing numbers of digits to recall.
Findings: Mean digit span 9.3 and mean letter span 7.3
MILLER
- Can recall 5 words as well as 5 words by chunking; receiving information from LTM increases STM capacity.
- Observed things come in 7s - suggested digit span was 7+-2
Overestimate? (COWAN found 4 chunks)
Duration of STM
PETERSON AND PETERSON
Procedure:
- 24 undergraduates each in 8 trials.
- Ppts given trigram to remember and a 3 digit number to count back from to prevent rehearsal.
- Stopped counting after period of time (3,6,9,12,15,18 secs).
Findings: Short duration; only 10% correct recall after 18 seconds
Duration of LTM
BAHRICK
Procedure:
- 392 ppts between 17 and 74 years from Ohio.
- Tested recall from high school year book:
1. Photo recognition: 50 pics some from yearbook
2. Free recall
Findings:
After 15 years - 90% PR and 60% FR
After 40 years - 70% PR and 30% FR
MSM separate stores SUPPORT
HM (case study)
- Normal STM (digit span) but no episodic memory from hours/mins before (LTM).
- Shows separate stores
MSM sensory register SUPPORT
SPERLING
Procedure:
- Partial report procedure
- Showed 3 lines of 4 letters too ppts for 50msec followed immediately by high, medium or low pitch (corresponding to line).
Findings: Average 3 correct responses, showing visual information available in sensory register.
MSM more than 1 type STM AGAINST
KF (case study - Shallice and Warrington)
- STM poor when being read aloud to but normal when read digits himself (different stores)
MSMmore than 1 type of LTM AGAINST
HM + WEARING (case studies)
- Poor LTM for events (episodic) but normal for meanings (semantic).
Types of LTM SUPPORT
HM + WEARING (case studies)
- Very poor episodic memory but unaffected semantic and procedural memory
- HM couldn’t remember stroking dog but understood what dog meant, Wearing could play piano.
TULVING
- Ppts did tasks involving different types of LTM during PET scan
- Procedural and semantic memory in prefrontal cortex, episodic in right prefrontal cortex.
WMM separate stores SUPPORT
KF (case study - Shallice and Warrington)
- Poor memory for verbal information but could process information normally
- Phonological loop damaged
- Shows separate visual and auditory stores
BADDELEY
- Dual task performance
- Doing 2 visual tasks more difficult than visual and verbal (tracing finger and picturing letter F vs tracing finger and saying days of week.
- Phonological loop and VSS separate
WMM phonological loop SUPPORT
BADDELEY
- Word length effect
- More difficult to remember longer words; finite space for maintenance rehearsal in articulately process.
- Word length effect disappears if given articulately suppression task tying up articulatory process.
Retroactive interference SUPPORT
MCGEOCH AND MCDONALD
Procedure:
- All had to learn 10 words until recall 100%
- Degree of similarity between first and second list dif for 6 groups:
1. Synonyms
2. Antonyms
3. Unrelated
4. Consonant syllables
5. 3-digit numbers
6. None
Findings: Synonyms for second list produced worst recall; retroactive interference worst when information most similar.
BADDELEY AND HITCH
- Rugby players asked to recall teams played that season.
- Recall worse if had played more games in meantime, unrelated to amount of time.
Interference SUPPORT
BURKE AND SKRULL
- Series of magazine adverts presented to ppts and asked to recall.
- Mixture of proactive and retroactive interference
State dependant forgetting SUPPORT
CARTER AND CASSADAY
Procedure: Learnt words either on/off antihistamine, (drug creates dif internal physiological state) and recall in 1 of 4 conditions.
- Learn on drug / recall on drug
- Learn on drug / recall off drug
- Learn off drug / recall on drug
- Learn off drug / recall off drug
Findings: Recall worse when different conditions; cues are absent b
Context dependant forgetting SUPPORT
GODDEN AND BADDELEY
Procedure: Divers had to learn list of words on/off land and recall them in 1 of 4 conditions.
- Learn on land / recall on land
- Learn on land / recall in water
- Learn in water / recall on land
- Learn on water / recall in water
Findings: Recall 40% worse in non-matching conditions because cues absent
Leading questions SUPPORT
LOFTUS AND PALMER
Procedure:
- Ppts watched films of car crash and asked questions. - In critical leading question ppts asked to describe how fast cars going.
- 5 groups, question had verb
‘contacted’, ‘bumped’, ‘hit,’collided’ or ‘smashed’
Findings: Mean estimate for ‘contacted’ was 31.8mph and for ‘smashed’ was 40.5mph.
Post-event discussion SUPPORT
GABBERT
Procedure:
- Pairs of ppts watched same crime but from different perspectives and saw different elements e.g. only 1 of pair could see title of book.
- Discussed in these pairs what had seen
- Completed recall test
Findings: 71% recalled aspects not seen in video (0% in control group with no discussion).
Anxiety negative SUPPORT
JOHNSON AND SCOTT
Procedure:
- Ppts led to believe participating in lab study and whilst in waiting room overheard argument. 2 conditions:
1. Low-anxiety condition: man walks through carrying pen with grease on hands.
2. High-anxiety condition: hear breaking glass and man walks through holding paper knife covered in blood.
Findings: Had to identify man that walked through from 50 photos. 49% of low anxiety condition managed and only 33% in high condition.
Anxiety positive SUPPORT
YUILLE AND CUTSHALL
Procedure:
- Quasi experiment
- 13 ppts witnessed shop owner shoot thief dead in Canada.
- Interviewed 4-5 months after + these compared to initial police interview; accuracy determined by number of correct details.
- Rated stress on 7 point scale + asked if any emotional problems since e.g. insomnia.
Findings: Little change in number or accuracy of details but those most stressed were 88% accurate and those less 73%
Yerkes-Dodson law SUPPORT
PARKER
- Interviews of those affected by Hurricane Andrew in US.
- Anxiety operationalised as damage suffered to homes.
- Moderate anxiety groups had best recall.
Enhanced cognitive interview SUPPORT/AGAINST
KOHNKEN
- Meta-analysis combining data from 50 studies, found ECI produced 81% more correct information than standard police interview.
- ECI 61% more incorrect information than standard police interview.
