All subjects · Psychology · WPS01 Social and Cognitive Psychology
The multi-store model of memory (Atkinson and Shiffrin, 1968)
Edexcel IAL Psychology · WPS01 Social and Cognitive Psychology
Cognitive psychology treats the mind as an information processor: input, processing, output. Spec 2.1.1 asks for the multi-store model, including encoding, storage, retrieval, capacity and duration.
Key terms
- Encoding: changing information into a form that can be stored (acoustic: sound; visual; semantic: meaning).
- Storage: holding information over time.
- Retrieval: getting stored information back out.
- Capacity: how much a store can hold.
- Duration: how long information lasts in a store.
The three stores
Information flows in one direction through three stores:
- Sensory register: holds input from each sense very briefly (well under 2 seconds); separate stores for each sense; very large capacity. Only information we pay attention to passes on.
- Short-term memory (STM): mainly acoustic encoding; capacity about 7 plus or minus 2 items (Miller, 1956), which can be increased by chunking; duration about 18 to 30 seconds without rehearsal. Maintenance rehearsal (repetition) keeps information in STM and transfers it to LTM.
- Long-term memory (LTM): mainly semantic encoding; capacity thought to be unlimited; duration potentially a lifetime. Information is retrieved back into STM when needed.
| Store | Encoding | Capacity | Duration |
|---|---|---|---|
| Sensory register | Modality specific | Very large | Under 2 seconds |
| STM | Mainly acoustic | 7 plus or minus 2 items | About 18 to 30 seconds |
| LTM | Mainly semantic | Unlimited | Up to a lifetime |
Supporting evidence
- Miller (1956): a review found people recall about 7 plus or minus 2 items (STM capacity).
- Peterson and Peterson (1959): participants saw consonant trigrams, then counted backwards in threes to stop rehearsal; recall fell to about 10% after 18 seconds (STM duration).
- Baddeley (1966): acoustically similar word lists were recalled worst immediately (STM is acoustic); semantically similar lists worst after a delay (LTM is semantic).
- Bahrick et al. (1975): about 90% recognition of classmates' faces after 15 years, and 70 to 80% after nearly 50 years (LTM duration).
- Glanzer and Cunitz (1966): a 30-second delay removed the recency effect but not the primacy effect, suggesting separate STM and LTM.
- HM: normal STM but unable to form new LTMs, suggesting separate stores.
Evaluation
Strengths: clear and testable; supported by lab research and brain-damage case studies (HM); influential, leading to the working memory model.
Weaknesses:
- STM is not a single store: patient KF (Shallice and Warrington, 1970) had poor verbal STM but better visual STM, which led to the working memory model (Baddeley and Hitch, 1974).
- LTM is not a single store either: episodic, semantic and procedural memory differ (Tulving, 1972).
- Rehearsal is not the only route to LTM; some things are remembered without rehearsal.
- Much evidence uses artificial tasks (trigrams, word lists), so ecological validity is low.
Worked example
A person reads a phone number once and keeps repeating it until they dial it. Use the multi-store model to explain why they forget it a minute later.
- Attention moved the number from the sensory register to STM.
- Maintenance rehearsal kept it in STM, but only while it was being repeated.
- Once rehearsal stopped it decayed within about 18 to 30 seconds and was displaced.
- It was not rehearsed enough to transfer to LTM.
Answer: It stayed in STM only while rehearsed; once rehearsal stopped it decayed or was displaced before transfer to LTM.
Check yourself
- What is the capacity of STM according to Miller (1956)?
- About 2 seconds
- About 7 plus or minus 2 items
- Unlimited
- About 4 chunks of sound only
Show answer
About 7 plus or minus 2 items. Miller's magic number seven, plus or minus two.
- How is information mainly encoded in LTM?
- Acoustically, by sound
- By the sense it came from
- Only visually
- Semantically, by meaning
Show answer
Semantically, by meaning. Baddeley (1966) found LTM confusions were based on meaning.
- Why did Peterson and Peterson make participants count backwards in threes?
- To test long-term memory
- To prevent rehearsal of the trigrams
- To measure the capacity of STM
- To make the task more realistic
Show answer
To prevent rehearsal of the trigrams. Counting stopped rehearsal so STM duration could be measured.
- According to the model, how does information move from STM to LTM?
- Through attention
- Through retrieval
- Through rehearsal
- Through the central executive
Show answer
Through rehearsal. Rehearsal transfers information; attention moves it from the sensory register to STM.
- Why does the case of KF challenge the multi-store model?
- His verbal STM was poor but visual STM was better, so STM may have separate parts
- He had perfect memory in all stores
- His LTM was destroyed but STM was perfect
- He could not attend to any information
Show answer
His verbal STM was poor but visual STM was better, so STM may have separate parts. Different STM abilities suggest STM is not a single store.
- Bahrick et al. found about 90% recognition after 15 years. If 50 photos were shown, about how many would be recognised?
Show answer
45 photos. 0.9 x 50 = 45 photos.
- Give one weakness of the multi-store model.
Show answer
It is too simple: STM and LTM are not single stores, for example LTM includes separate episodic, semantic and procedural memory.. STM and LTM each seem to have several components; rehearsal is not the only route.
Sources
- Pearson Edexcel International Advanced Subsidiary/Advanced Level in Psychology specification (issue January 2026)
- Pearson International A Level Psychology component guide: WPS01
- Simply Psychology: Multi-store memory model
- IAL Psychology revision site: The multi-store model of memory
Last updated 8 November 2026. Spotted a mistake? Tell us.