Cognitive Psychology: Memory Models, Capacity, Duration, and Coding

Fundamental Memory Concepts

• Capacity:
  • The total amount of information that a given memory store can hold at any single time.
• Duration:
  • The length of time that information can remain accessible within a memory store before decaying or being lost.
• Coding:
  • The format or mental representation in which information is processed and held in a memory store (e.g., acoustically by sound, visually by image, or semantically by meaning).
• Chunking:
  • The cognitive process of grouping individual units of information into larger, meaningful clusters or units to extend the effective capacity of short-term memory (STM).
  • Example: Remembering the $11$-digit phone sequence $07813999667$ as three distinct chunks—$07813$, $999$, and $667$—thereby occupying only three functional spaces in STM rather than eleven independent slots.

Empirical Research on Memory Capacity, Coding, and Duration

• Capacity of Short-Term Memory (STM) — Jacobs (1887):

  • Aim: To determine the storage capacity limits of human short-term memory.
  • Procedure:
  • Developed and administered the original digit span test.
  • Participants were presented with increasing sequences of numbers or letters and required to recall them immediately in the exact sequence of presentation.
  • The length of the sequences increased systematically until the participant could no longer recall the sequence without errors.
  • Results:
  • Mean digit span: $9.3$ items.
  • Mean letter span: $7.3$ items.
  • Conclusions:
  • STM possesses a strictly limited capacity.
  • The type of stimulus material influences retention span, as individuals routinely exhibit a larger span for numeric digits than for alphabetic letters.

• Coding in Short-Term Memory (STM) — Conrad (1964):

  • Aim: To examine whether STM encodes information acoustically or visually.
  • Procedure:
  • Displayed visual letter sequences comprising six letters at a presentation speed of one letter every $0.75\,\text{seconds}$.
  • Target letters selected were B, C, F, M, N, P, S, T, V, and X.
  • Participants were instructed to record the letters in the exact serial order immediately following presentation.
  • Conrad classified and analyzed the specific categories of recall errors rather than merely tabulating aggregate recall scores.
  • Results:
  • Participants routinely made acoustic confusion errors, substituting target letters with letters that sounded similar rather than letters that possessed visual similarity (e.g., substituting B with V, or F with S).
  • Conclusions:
  • STM relies primarily on acoustic coding (sound-based representation), even when stimuli are delivered through visual channels.

• Duration of Short-Term Memory (STM) — Peterson and Peterson (1959):

  • Aim: To quantify the retention duration of STM when rehearsal is actively blocked.
  • Procedure:
  • Participants were presented with three-letter nonsense syllables known as consonant trigrams (e.g., XQJ).
  • Immediately following the presentation, participants were given a starting number and instructed to count backwards in intervals of three to block maintenance rehearsal.
  • Participants were stopped and asked to recall the trigram after variable retention intervals.
  • Results:
  • Recall performance deteriorated rapidly over the delay period; by approximately $18\,\text{seconds}$, the vast majority of participants could no longer reproduce the trigram correctly.
  • Conclusions:
  • Without active maintenance rehearsal, information in STM has a duration of approximately $18\text{--}30\,\text{seconds}$.

• Duration of Long-Term Memory (LTM) — Bahrick et al. (1975):

  • Aim: To evaluate the persistence and durability of long-term memory over extended chronological intervals.
  • Procedure:
  • Tested participants on their memory of former high school peers using original school yearbooks.
  • Assessment included recognition of classmates' faces from photos and verbal recall of classmates' names.
  • Results:
  • Face recognition accuracy was approximately $90\%$ among individuals tested $15\,\text{years}$ post-graduation.
  • Face recognition accuracy remained at approximately $70\%$ after a retention interval of roughly $48\,\text{years}$.
  • Conclusions:
  • Information stored in LTM can endure across multiple decades and can persist for an entire human lifetime.

Methodological Evaluation of Research into Memory Stores

• Methodological Strengths — High Experimental Control:

  • Laboratory paradigms enabled precise manipulation of independent variables and isolation of target constructs.
  • Conrad (1964) established standardized conditions by presenting each letter for exactly $0.75\,\text{seconds}$ and presenting uniform letter arrays across all participants.
  • Rigorous standardization diminished extraneous variables, providing confidence that confusion errors reflected intrinsic acoustic coding mechanisms rather than procedural discrepancies.
  • This high level of experimental control directly reinforces the internal validity of the experimental findings.

• Methodological Weaknesses — Artificial Stimulus Materials:

  • Many foundational memory experiments relied on artificial, decontextualized materials, including randomized numbers, arbitrary letter strings, and artificial consonant trigrams.
  • These stimuli fail to capture the meaningful, contextual, and emotional characteristics of information encountered and retained in daily human activity.
  • The use of artificial tasks compromises external (ecological) validity, as lab-based memorization tasks may not accurately reflect naturalistic human memory function in real-world environments.

The Multi-Store Model of Memory (MSM)

• Theoretical Overview — Atkinson and Shiffrin (1968):

  • Conceptualizes memory as a flow of information through three discrete, structural storage systems:
  • Sensory Register (SR)
  • Short-Term Memory (STM)
  • Long-Term Memory (LTM)
  • Information progresses linearly from one store to the next via regulatory cognitive operations, specifically attention and rehearsal.

• Sensory Register (SR):

  • Entry Point: Receives raw, unencoded stimuli from the external environment via the sense organs.
  • Capacity: Very large, encompassing the complete perceptual field.
  • Duration: Extremely brief, lasting around half a second ($0.5\,\text{seconds}$) for most sensory modalities.
  • Coding: Modality-specific (information remains encoded according to the sense that registered it; e.g., acoustic coding for auditory input, visual coding for iconic input).
  • Transfer Mechanism: Unless immediate attention is directed to the incoming sensory signal, the information decays rapidly and is permanently forgotten.

• Short-Term Memory (STM) Store:

  • Transfer Mechanism: Inputs that receive focused attention in the Sensory Register transfer directly to STM for temporary holding.
  • Capacity: Limited to roughly $5\text{--}9\,\text{items}$.
  • Capacity can be extended through chunking (e.g., condensing the sequence $343565787$ into $343$, $565$, and $787$, reducing nine separate items into three recallable units).
  • Duration: Approximately $18\text{--}30\,\text{seconds}$ in the absence of active maintenance rehearsal.
  • Coding: Primarily acoustic, though visual and semantic encoding can also occur.
  • Function in MSM: Acts as a working buffer; information is maintained through repetitive maintenance rehearsal or retrieved from LTM back into STM to guide active behavior.

• Long-Term Memory (LTM) Store:

  • Transfer Mechanism: Information is consolidated from STM into LTM through repeated maintenance rehearsal.
  • Capacity: Potentially unlimited.
  • Duration: Potentially permanent, lasting for decades or an entire lifetime.
  • Coding: Predominantly semantic, organizing and storing information on the basis of meaning and conceptual connections.
  • Retrieval: Stored representations must be transferred from LTM back into STM to be accessed and reported.

Critical Evaluation of the Multi-Store Model (AO3)

• Empirical Support: Neuropsychological Dissociation (Case of KF):

  • Shallice and Warrington (1970) documented the clinical case of patient KF following a severe motorcycle accident.
  • KF presented with profound impairment to short-term memory function, exhibiting an auditory digit span limited to approximately $2\,\text{items}$.
  • In contrast, KF's long-term memory capacity and recall remained largely intact.
  • Implication: This dissociation supports the structural distinction posited by the Multi-Store Model, confirming that STM and LTM represent physiologically distinct memory systems. If memory were an indivisible, unitary store, traumatic brain injury would affect all temporal spans equally.

• Theoretical Limitation: Oversimplification of Memory Stores:

  • The MSM treats both STM and LTM as homogeneous, unitary compartments, an assumption contradicted by subsequent clinical and experimental research.
  • Tulving (1985) demonstrated that LTM consists of qualitatively distinct sub-systems:
  • Episodic Memory: Storage of personally experienced events, autobiographical episodes, and temporal-spatial contexts.
  • Semantic Memory: Generalized storage of factual information, concepts, and abstract world knowledge.
  • Procedural Memory: Non-declarative storage of motor actions, perceptual-motor skills, and automated habits.
  • Implication: Describing LTM as a solitary store fails to capture the complexity and neural dissociation among these distinct mnemonic systems.

• Theoretical Limitation: Inadequacy of Maintenance Rehearsal as the Sole Transfer Route:

  • Atkinson and Shiffrin posited that the volume of maintenance rehearsal directly determines the likelihood of transfer from STM into LTM.
  • Hyde and Jenkins (1973) disproved this assertion by demonstrating that participants exposed to incidental learning tasks achieved robust long-term retention without engaging in deliberate or repeated rehearsal.
  • The depth and qualitative nature of cognitive processing during encoding governs memory retention far more than the simple quantity of rehearsal.
  • This evidence aligns with Craik and Lockhart's (1972) Levels of Processing framework, directly undermining the MSM claim that rote rehearsal is the primary gateway to long-term memory storage.