Short-Term Memory vs Working Memory: What's the Difference and Why It Matters

Early memory research in the 1950s and 60s proposed a simple two-stage model: a short-term store held a small amount of information for seconds, and a long-term store held vast amounts indefinitely. George Miller's famous 1956 paper established that the short-term store held approximately 7 ± 2 items — a limit that became embedded in popular psychology.

The short-term store was conceived as passive: information arrived, was held briefly, and either decayed or was transferred to long-term memory through rehearsal. This model could not easily explain how people performed complex tasks like mental arithmetic, language comprehension, or reasoning, which require not just holding information but actively manipulating it.

In 1974, Alan Baddeley and Graham Hitch proposed a working memory model that replaced the passive short-term store with an active, multi-component system. The central executive (a limited-capacity attentional controller) coordinates two slave systems: the phonological loop (rehearsing verbal and acoustic information) and the visuospatial sketchpad (maintaining and manipulating visual and spatial information). A later addition, the episodic buffer, integrates information from both slaves and from long-term memory.

The key distinction is that working memory is not just storage — it is storage plus processing. When you hold a phone number in mind while deciding whether to call, you are using the phonological loop. When you mentally rotate a shape to check if it matches a target, you are using the visuospatial sketchpad. When you track which moves you have already tried in a puzzle while planning the next one, you are using the central executive.

Working memory capacity — not short-term memory span — predicts academic achievement, reading comprehension, mathematical ability, and fluid intelligence. The reason is that virtually all complex cognitive tasks require active manipulation of information, not just passive storage. Understanding a sentence requires holding early words active while processing later ones and integrating their meaning. Solving a multi-step problem requires tracking intermediate results while applying the next operation.

People with high working memory capacity are better at ignoring distractions, switching between tasks, and maintaining goal-directed behaviour under cognitive load. These are executive function skills, which is why working memory is considered both a component and a strong predictor of executive function more broadly.

Number Memory directly measures phonological loop capacity: a digit string is displayed, you rehearse it, and recall it without any competing task. This is close to the classical short-term memory span paradigm. Visual Memory and Pattern Recall target the visuospatial sketchpad: a spatial pattern must be held in mind and reproduced after it disappears.

The more demanding working memory games add a processing component on top of storage. In Arrow Out, you must hold the current board state while mentally simulating future moves. In Sudoku, you track which digits remain available in each region while deducing where to place the current one. Color Memory adds processing difficulty by requiring you to reconstruct a stored colour rather than simply recognise it — a finer-grained visuospatial working memory task.

Working memory capacity is partially heritable and shows a gradual developmental increase through childhood and adolescence, peaking in the mid-20s. In adulthood, it can be temporarily degraded by sleep deprivation, stress, and distraction — and temporarily enhanced by physical exercise and sufficient sleep.

Training studies show that practising working memory tasks improves performance on those specific tasks reliably. Transfer to other tasks is more limited: there is good evidence for near transfer (to untrained tasks using the same component, e.g., other visuospatial working memory tasks) and weaker evidence for far transfer (to general fluid intelligence). The most honest summary is that consistent mental challenge maintains working memory function better than it expands its ceiling.