Models of memory ERQ

Cognitive models in memory research

To represent and explain the structure and processes of memory using simplified frameworks based on experimental data.

Multi-Store Model (MSM)

Atkinson & Shiffrin (1968)

Working Memory Model (WMM)

Baddeley & Hitch (1974)

Main components of the Multi-Store Model

Sensory Memory, Short-Term Memory (STM), Long-Term Memory (LTM); Memory transfers through attention and rehearsal.

Aim of Glanzer & Cunitz (1966)

To test whether STM and LTM are separate stores by examining the serial position effect.

Procedure of Glanzer & Cunitz study

Experiment 1: 240 military personnel presented with list of 20 common nouns, varying the amount of times the words were repeated and how long they were shown on the screen for. Then asked to write down nouns

Experiment 2: 47 military personnel presented with list of 15 common nouns and then counted backwards (distraction task to prevent rehearsal) before being asked to write down the nouns

Key findings of Glanzer & Cunitz study

Experiment 1: Spacing and amount of repetition mainly affected words in beginning of sequence but not so much later ones (primacy effect)

Experiment 2: Immediate recall: primacy and recency effects observed; Delayed recall: recency effect disappeared; Suggests separate STM and LTM stores.

Conclusion of Glanzer & Cunitz study

Supports MSM's idea of separate memory stores (STM vs. LTM).

Components of the Working Memory Model

Central Executive, Phonological Loop, Visuospatial Sketchpad, Episodic Buffer (added later).

Aim of Robbins et al. (1996)

To test if different components of working memory could be selectively overloaded.

Procedure of Robbins study

male chess players from Cambridge, repeated measures. 2 conditions: in both, they view the arrangement of chess pieces and recreated them on a next board. 1. verbal interference - repeat a word as they commit to memory 2. visuo-spatial interference - type a word while committing to memory

Findings of Robbins study

those in the verbal interference condition were able to recall the arrangement in ~65% trials and those in the visuo-spatial interference were correct in only about 15%.

Conclusion of Robbins study

Supports the WMM by showing the independence of subsystems and the role of the central executive.

Limitation of the MSM addressed by WMM

MSM oversimplifies STM as a single store; WMM explains STM as a multi-component, active system.

Contribution of MSM to understanding memory

Shows how memory is transferred and stored.

Contribution of WMM to understanding memory

Shows how information is processed and manipulated in STM.

Overall conclusion regarding the two models

Both models offer valuable but different perspectives—MSM emphasizes structure; WMM emphasizes function and processing.

Study supporting MSM

Glanzer & Cunitz → supports MSM.

Study supporting WMM

Robbins et al. → supports WMM.

Strengths of Glanzer & Kunitz

Controlled lab experiment:

• High internal validity due to control over variables like word presentation time, list length, and recall timing

• Hence also replicable

Limitations of Glanzer & Kunitz

• Low ecological validity:

  • Artificial task (word list recall) does not reflect real-world memory use.

• Sample bias:

  • Often conducted with military personnel or students, limiting generalizability.

Strengths of Robbins et al

Sample consisted of both professional and amateur chess players - shows that it wasn’t a difference in skill that caused the difference between the two conditions but slave systems being overloaded

Supports the idea of modality-specific interference

• The study shows that different cognitive tasks interfere with different memory systems (e.g., visuospatial vs. phonological), supporting modality-specific processing.

• This is a strength over the Multi-Store Model, which doesn’t explain how STM processes different types of information.

Repeated measures design ensured no participant-dependent factors influenced the results

Limitations of Robbins et al

Task demands may cause distraction rather than system overload

• The visuospatial task (random number generation) is cognitively demanding — it's hard to tell if poor performance is due to system-specific interference or simply cognitive overload or distraction.