Short-Term Memory is a limited-capacity cognitive system that holds small amounts of information in conscious awareness for roughly 15 to 30 seconds without active rehearsal. George Miller’s 1956 research proposed a storage ceiling of seven plus or minus two chunks; Cowan’s reanalysis of decades of memory-span data revised this estimate to approximately four discrete items.
The term is often used interchangeably with working memory, though formally short-term memory refers to passive storage only, while working memory encompasses active manipulation of that stored content.
The capacity of short-term memory was established by Miller, whose 1956 experiments showed subjects reliably hold seven plus or minus two chunks of information at once 1. A chunk is any meaningful unit the learner can recognise as a single item, regardless of its internal complexity: a word, a digit cluster, a familiar acronym. Cowan reanalysed decades of memory-span data and argued the true limit, once rehearsal and chunking artefacts are controlled, sits closer to four items 2. The 7±2 figure persists in popular accounts; the four-item estimate is now the scientific consensus.
Baddeley’s multicomponent model situates short-term storage within two subordinate systems: a phonological loop that maintains verbal and auditory material through articulatory rehearsal, and a visuospatial sketchpad that holds visual and spatial representations 3. Both feed into a central executive that allocates attentional resources, and an episodic buffer that integrates incoming material with content already held in long-term memory. Without active rehearsal, phonological traces decay in approximately 15 to 20 seconds, consistent across articulatory suppression studies 3.
A new recruit given a ten-digit radio frequency struggles to retain it while simultaneously following a briefing; an experienced operator, who has internalised frequency patterns as single chunks, carries the same number effortlessly alongside multiple other inputs. The novice holds ten items; the expert holds three. The channel’s information load has not changed; the expert’s short-term memory simply processes it differently.
Short-term memory capacity is not fixed in practice: expertise reorganises incoming information into fewer, richer chunks, expanding effective capacity without changing the underlying limit.
The capacity ceiling of short-term memory is the primary bottleneck in skill acquisition. Novices cannot yet organise incoming information into efficient chunks, so identical material places heavier demands on their short-term memory than on an expert’s 4. Cognitive load theory, built directly on these limits, predicts that instruction exceeding available short-term memory capacity impairs learning; two decades of controlled experiments confirm that reducing extraneous load improves both knowledge acquisition and transfer to new contexts 4.
Information held in short-term memory transfers to long-term memory through elaborative encoding and rehearsal; without this transfer, newly encountered material is typically lost within 30 seconds 3. Practical applications follow directly: dual-channel instruction, which presents verbal content as audio alongside corresponding visuals, distributes the load across the phonological loop and visuospatial sketchpad simultaneously, raising effective capacity for complex material 4.
Without active rehearsal, information held in short-term memory typically decays within 15 to 20 seconds. This decay rate is consistent across articulatory suppression studies and explains why phone numbers fade almost immediately if you do not repeat them. Rehearsal, either covert or spoken, resets this clock and keeps the trace active.
Short-term memory and working memory are related but distinct. Short-term memory refers to the passive storage of a small number of items over a brief period. Working memory, in Baddeley’s model, extends this to include active manipulation of stored material via the phonological loop, visuospatial sketchpad, and central executive.
Miller’s classic 1956 experiments found a limit of seven plus or minus two chunks. Cowan’s later reanalysis, controlling for rehearsal and chunking artefacts, placed the true limit at approximately four items. The four-item estimate is now the scientific consensus, while the 7±2 figure remains the more widely cited popular statistic.
The most reliable strategy is chunking: grouping individual items into meaningful units reduces the number of items your short-term memory must hold simultaneously. Dual-channel instruction (audio for verbal content, visuals for diagrammatic content) distributes load across both the phonological loop and visuospatial sketchpad. Spacing rehearsal sessions also consolidates material into long-term storage.
Why Incompetence Feels Like Competence: The Dunning-Kruger Effect Examined
Applied Flow Protocols: Domain-Specific Systems for Reliable Peak Performance
Burnout Test: Where Are You on the Burnout Spectrum Right Now?
90-Day Sleep Optimisation Protocol: Rebuild Your Recovery From the Ground Up
Digital Detox Science: What Actually Happens When You Block Algorithmic Feeds
The Psychology of Power: What Happens to the Brain When You Gain Authority
Cognitive Fuel: The Evidence-Based Nutritional Framework for Brain Performance
Network Intelligence: The Science of Strategic Relationship Building for Career Growth
The 90-Day Kickstarter Protocol
Your day-by-day reset for sleep, stress & energy · PDF