Lesson 7 of 11 · 14 min
Comparing the three models of memory
Compare the multi-store, working memory and levels of processing models on their structure, view of short-term memory, role of rehearsal and evidence, and discuss how new evidence contested and refined them.
What you will be able to do
- Compare the three models of memory on the same features, with similarities and differences.
- Explain a finding that one model explains and another cannot.
- Discuss how models of memory have been contested, refined and replaced as new evidence appeared.
- Write a comparison that would earn full marks in the external examination.
Three models, three different questions. The multi-store model asks where information is held. The working memory model asks how short-term memory does its work. The levels of processing model asks what happens to information at encoding. They are not simply right or wrong: each explains some findings well and others poorly, and each was built to fix a weakness in the one before.
| Feature | Multi-store (Atkinson & Shiffrin | Working memory (Baddeley & Hitch | Levels of processing (Craik & Lockhart | |||
|---|---|---|---|---|---|---|
| What memory is | Three separate stores in a fixed sequence | An active workspace (STM) with several components, beside LTM | A by-product of processing; no stores | |||
| Short-term memory | One store: about 7 ± 2 items, about 18–30 s | Central executive, phonological loop, visuospatial sketchpad, episodic buffer | Not a separate store; shallow processing gives brief memory | |||
| How information reaches LTM | Rehearsal: more rehearsal, more transfer | Not its focus; the episodic buffer links working memory with LTM | Depth of encoding: semantic processing and elaborative rehearsal | |||
| Key evidence | Serial position effect; HM; Peterson & Peterson (1959) | Dual-task studies; word-length effect; KF | Hyde & Jenkins (1973); Craik & Tulving (1975) | |||
| Main weakness | STM is not unitary; rehearsal is not the only route to LTM | Central executive vague; little about LTM | Depth cannot be measured independently (circular) |
Exam tip
How to compare
"Compare" asks for similarities and differences. Take one feature at a time and say what each model claims about it, joined by whereas, both or unlike:
Both the multi-store and working memory models describe a limited short-term memory, but the multi-store model treats it as a single store whereas the working memory model divides it into components that can work at the same time.
Two separate paragraphs, one on each model, describe them; they do not compare them.
Match each claim to the model that makes it.
Models contested, refined and replaced. Psychology's models of memory changed because evidence the old model could not explain kept appearing.
- The multi-store model was contested by evidence that STM is not unitary (KF; dual-task studies). Rather than patch it, Baddeley and Hitch replaced its short-term store with working memory. The multi-store model's broad distinction between short-term and long-term memory survived inside the new model.
- The multi-store model's account of rehearsal was contested by evidence that the kind of processing matters more than the amount of repetition. Craik and Lockhart offered levels of processing as an alternative framework.
- Levels of processing was refined after Craik and Tulving (1975) showed that elaboration and the "yes" effect matter, and after Morris, Bransford and Franks (1977) showed that the match between encoding and the test matters.
- Working memory was refined in the 1980s (the phonological store and articulatory control process) and in 2000, when the episodic buffer was added to explain how working memory uses LTM and combines information into episodes.
How new evidence led from one model to the next.
Worked example
Compare the multi-store and working memory models
Compare the multi-store model and the working memory model's accounts of short-term memory. (4 marks)
Step 1: A similarity
Both models describe a short-term system with a limited capacity that holds information we are consciously using, separate from long-term memory.
Step 2: A difference in structure
The multi-store model treats STM as a single, unitary store, whereas the working memory model divides it into components: the central executive, phonological loop, visuospatial sketchpad and episodic buffer.
Step 3: A difference in function
In the multi-store model STM is mainly a passive holding store where information is rehearsed for transfer to LTM, whereas working memory is an active workspace where information is manipulated during reasoning, reading and calculation.
Step 4: A difference in evidence
Unlike the multi-store model, the working memory model can explain why two tasks can be done together when they use different components (dual-task studies) and why KF lost verbal but not visual STM.
Try it yourself first, then check the working one step at a time.
Answer
Both describe a limited short-term system separate from LTM. The multi-store model treats STM as one passive store for rehearsal, whereas working memory divides it into active components (central executive, phonological loop, visuospatial sketchpad, episodic buffer), which lets it explain dual-task findings and KF that the multi-store model cannot.
Method
Choose features: structure, function, evidence.
For each, state both models' positions in one sentence with "whereas", "both" or "unlike".
Watch out for
Describing each model in its own paragraph without linking them.
A student can hold a phone number in mind while picturing the route to a friend's house, but cannot hold the phone number while reciting a poem. Explain how the working memory model accounts for this, and why the multi-store model cannot.
Note
Why the evidence is open to interpretation
Memory cannot be observed directly. Researchers infer it from what people recall, recognise or relearn, usually with artificial material (trigrams, word lists) in a laboratory. The same result can fit more than one model: the fall in recall in Peterson and Peterson (1959) fits decay or interference; the serial position curve fits two stores or two kinds of processing. Case studies such as HM and KF are rich but rare, and a person's brain injury is never as neat as a model's box. So memory models are best seen as useful descriptions that are kept, refined or replaced as evidence accumulates.
Summary
Multi-store: three stores in sequence; rehearsal transfers to LTM. Working memory: STM as components working together. Levels of processing: memory depends on depth of encoding.
Compare feature by feature: what memory is, what STM is, how information reaches LTM, the evidence and the weaknesses.
Working memory has greater explanatory power than the multi-store STM: it explains dual-task performance, the word-length effect and KF.
Levels of processing explains memory for things never rehearsed, but depth cannot be measured independently.
Each model was contested by evidence and refined or replaced; the data are often open to more than one interpretation.
Finished the lesson?
Try the 2 quick checks above and the lesson completes itself, or mark it complete now.