The Multi-Store Model of Memory

The Multi-Store Model (MSM) was proposed by Atkinson & Shiffrin (1968) and was the first comprehensive model of how memory works. It describes memory as a linear flow of information through three distinct stores: the sensory register, short-term memory (STM), and long-term memory (LTM). Despite its limitations, the MSM remains a foundational model in cognitive psychology and is a core component of the AQA A-Level specification.

Key Definition: The Multi-Store Model is a representation of how memory works in terms of three stores (sensory register, STM, LTM) and the processes that transfer information between them (attention, encoding, rehearsal, retrieval).

---

The Three Stores

graph LR
    A["Sensory Register"] -->|Attention| B["Short-Term Memory (STM)"]
    B -->|Maintenance Rehearsal| B
    B -->|Elaborative Rehearsal| C["Long-Term Memory (LTM)"]
    C -->|Retrieval| B
    A -->|Not attended to| D["Information lost"]
    B -->|Not rehearsed| E["Information lost via decay/displacement"]

1. Sensory Register (SR)

The sensory register is the first store. All incoming sensory information enters here.

Feature	Detail
Duration	Very brief — less than half a second for visual (iconic) and 2–4 seconds for auditory (echoic)
Capacity	Very large — receives all sensory input simultaneously
Encoding	Modality-specific — stored in the form it was received (visual, auditory, haptic, etc.)

• Iconic memory (visual) was demonstrated by Sperling (1960): participants were shown a grid of 12 letters for 50 milliseconds. When asked to recall all letters (whole report), they recalled about 4–5. But when a tone indicated which row to recall (partial report), they could recall 3 out of 4 from any row, suggesting all 12 were briefly stored.
• Information passes from the SR to STM through attention. Information that is not attended to is lost.

2. Short-Term Memory (STM)

STM is the second store, where information is held temporarily for immediate use.

Feature	Detail	Key Research
Capacity	Limited to approximately 7 ± 2 items	Miller (1956) — "The magical number seven, plus or minus two." Miller also demonstrated that chunking can increase the effective capacity by grouping individual items into meaningful units.
Duration	18–30 seconds without rehearsal	Peterson & Peterson (1959) — gave participants a consonant trigram (e.g., BKG) and asked them to count backwards in threes from a specified number (to prevent rehearsal). After 3 seconds, recall was about 80%; after 18 seconds, only about 3%. This shows STM duration is very short without maintenance rehearsal.
Encoding	Primarily acoustic (sound-based)	Baddeley (1966) — gave participants lists of acoustically similar words (e.g., cat, cab, can) and acoustically dissimilar words. Recall of acoustically similar words from STM was significantly worse, suggesting STM relies on acoustic encoding.

Key Definition: Maintenance rehearsal is the process of repeating information over and over to keep it in STM. According to the MSM, prolonged maintenance rehearsal leads to transfer to LTM.

Feature	Detail	Key Research
Digit span	A method of measuring STM capacity	Jacobs (1887) — found the average digit span was 9.3 items for digits and 7.3 items for letters. Digits are easier because there are only 10 possible digits (0–9) compared with 26 letters.

3. Long-Term Memory (LTM)

LTM is the third store, where information can be held for very long periods.

Feature	Detail	Key Research
Capacity	Potentially unlimited	No study has been able to measure the upper limit of LTM capacity
Duration	Up to a lifetime	Bahrick et al. (1975) — tested recall of high school classmates' names and faces. After 15 years, participants could identify 90% of names and faces from photographs. Even after 48 years, recognition was about 70%. Free recall was worse (about 30% after 48 years), but this still demonstrates very long-term storage.
Encoding	Primarily semantic (meaning-based)	Baddeley (1966) — found that recall of semantically similar words (e.g., big, large, great) from LTM was significantly impaired compared with semantically dissimilar words, suggesting LTM relies on semantic encoding.

---

Processes in the MSM

Process	Description
Attention	Transfers information from the sensory register to STM; without attention, information is lost
Maintenance rehearsal	Repeating information to keep it in STM; prolonged rehearsal transfers it to LTM
Retrieval	Bringing information back from LTM into STM for conscious use

Chunking

Key Definition: Chunking is the process of combining individual items into larger, meaningful units (chunks) to increase the effective capacity of STM.

Miller (1956) demonstrated that while STM can hold approximately 7 items, the size of each item can vary. For example, the sequence "F-B-I-C-I-A-B-B-C" is 9 individual letters (exceeding STM capacity), but can be chunked into three meaningful units: "FBI," "CIA," "BBC" — well within STM capacity. Chunking relies on existing knowledge in LTM to create meaningful groups, which is why experts in a field can remember more domain-specific information than novices.

---

Evaluation of the Multi-Store Model (AO3)

Strengths