Memory detailed chatGPT

Multi-Store Model (MSM)

Linear model with 3 stores: sensory register → STM → LTM. Info passes via attention and rehearsal.

Sensory register

Stores sensory info for <0.5 seconds, very large capacity, separate stores for each sense.

STM coding

Acoustic (sound-based).

STM capacity

7±2 items (Miller).

STM duration

18-30 seconds.

LTM coding

Semantic (meaning-based).

LTM capacity

Unlimited.

LTM duration

Potentially lifelong.

Rehearsal

Maintenance rehearsal keeps info in STM; prolonged rehearsal transfers to LTM.

MSM strength

Peterson & Peterson (1959) supports STM duration (~18 seconds).

MSM limitation

Oversimplified - LTM has different types (episodic, semantic, procedural).

MSM limitation (ecological validity)

Uses artificial stimuli (e.g. trigrams), so low ecological validity.

Coding

The format information is stored in (acoustic, visual, semantic).

Capacity

The amount of information memory can hold.

Duration

How long information can be stored.

Baddeley (1966) coding study

STM = acoustic, LTM = semantic (based on recall errors).

Jacobs (1887) capacity study

Digit span 9.3, letter span 7.3 → STM limited capacity.

Miller (1956)

STM capacity = 7±2 chunks; chunking increases capacity.

Peterson & Peterson (1959) duration study

Trigrams + counting backwards → STM lasts ~18 seconds.

Bahrick et al (1975) duration study

Recall of classmates up to 46 years → LTM duration very long.

Capacity research limitation

Lack of standardisation (Jacobs).

Bahrick strength

High ecological validity (real-life memories).

Working Memory Model (WMM)

STM is made of multiple components, not one unitary store.

Central executive

Allocates attention/resources to slave systems; limited capacity.

Phonological loop

Processes verbal/auditory info (articulatory process + phonological store).

Visuo-spatial sketchpad

Stores visual/spatial info; capacity ~4-5 items.

Episodic buffer

Integrates info and links STM to LTM.

WMM strength (KF)

KF had poor verbal STM but good visual STM → separate stores supported.

WMM strength (dual-task)

Performance worse when tasks use same store → limited capacity supported.

WMM limitation

Central executive is vague and not clearly defined.

WMM limitation (ecological validity)

Lab tasks unrealistic (e.g. repeating words).

Interference

Forgetting occurs when memories compete.

Proactive interference

Old info interferes with new learning.

Retroactive interference

New info interferes with recall of old info.

When is interference strongest?

When memories are similar.

Baddeley & Hitch (1977)

Rugby players recall worse with more games → retroactive interference.

Keppel & Underwood (1962)

First trigrams recalled best → proactive interference.

Interference strength

Supported by controlled lab studies.

Interference limitation

Artificial tasks → low ecological validity.

Interference limitation

Cannot explain forgetting when info is dissimilar.

Retrieval failure

Forgetting due to absence of cues.

Encoding specificity principle (ESP)

Recall best when encoding cues = retrieval cues.

Context-dependent forgetting

External environment mismatch reduces recall.

State-dependent forgetting

Internal state mismatch reduces recall.

Godden & Baddeley (1975)

Divers recall better in matching environments.

Goodwin et al (1969)

Recall better drunk-drunk or sober-sober.

Tulving & Pearlstone (1966)

Recall improves with cues (40% → 60%).

Retrieval failure strength

Strong lab support.

Retrieval failure limitation

Cues not as extreme in real life → low ecological validity.

Retrieval failure limitation

ESP is circular explanation.

Leading questions

Questions that suggest a desired answer.

Loftus & Palmer (1974) procedure

Watched car crash, estimated speed with different verbs.

Loftus & Palmer findings

"Smashed" → higher speed estimates than "contacted".

Loftus & Palmer follow-up

"Smashed" group more likely to report false broken glass.

Substitution explanation

Leading questions change memory itself.

Leading questions limitation

Lab study → low ecological validity.

Leading questions limitation

Sample = American students → low population validity.

Leading questions strength

Real-world application to police interviews.

Post-event discussion (PED)

Co-witness discussion leading to memory distortion.

Gabbert et al (2003) procedure

Different videos → discussion → recall test.

Gabbert findings

71% recalled false info; 60% falsely said guilty.

PED explanation

Memory contamination or conformity.

PED limitation

Lab study → low ecological validity.

PED limitation

Mechanism unclear (conformity vs memory change).

PED strength

Police should interview witnesses separately.

Anxiety effect on EWT

Can increase OR decrease accuracy.

Johnson & Scott (1976)

Weapon focus reduces identification accuracy (33% vs 49%).

Weapon focus

Attention on weapon reduces recall of perpetrator.

Yuille & Cutshall (1986)

Real-life shooting → high anxiety = more accurate recall.

Yerkes-Dodson Law

Inverted U: moderate anxiety = best performance.

Anxiety strength

Real-life research → high ecological validity.

Anxiety limitation

Ethical issues (distress).

Anxiety limitation

Contradictory findings.

Cognitive interview

Method to improve EWT using retrieval cues.

Report everything

Include all details → triggers memories.

Reinstate context

Mentally recreate environment.

Change order

Reduces schema influence.

Change perspective

Recall from different viewpoints.

Cognitive interview strength (Fisher et al.)

Trained detectives → more detailed recall.

Cognitive interview limitation (Köhnken et al.)

Increases incorrect recall too.

Cognitive interview limitation

Time-consuming and requires training.

Cognitive interview strength

Useful in real police work.