Short-Term Memory In Psychology

Short-term memory has three key aspects:

1. Limited Capacity: Only about seven items can be stored at a time.
2. Limited Duration: Storage is fragile, and information can be lost through distraction or the passage of time.
3. Encoding: Information is stored primarily by sound, even when it enters through the eyes.

Capacity: Magic Number 7

The capacity of short-term memory is limited. Miller (1956) proposed a classic theory: the average person can hold about seven objects in short-term memory. The true number varies by two either way.

Miller thought this was because short-term memory has a fixed number of “slots.” Each slot stores one item.

However, Miller didn’t specify how much information each slot could hold. This is the key to chunking: grouping single items into a larger, meaningful unit lets far more information fit into the same seven slots.

Chunking works like this:

The digits “1”, “0”, “6”, “6” become one chunk: the date 1066. Chunking works the same way for phone numbers. Splitting a ten-digit number into a few short blocks makes it memorable. This is why an expert can hold far more of their subject than a novice can.

Miller’s theory is supported by earlier evidence from Jacobs (1887), using the digit-span procedure. Jacobs presented digit or letter sequences of increasing length. He measured the longest string each person could recall correctly. He excluded “w” and “7” because both have two syllables.

People recalled numbers more easily than letters: the average span was 9.3 for digits but only 7.3 for letters.

However, the nature of the items (e.g., simple versus complex) and individual differences can influence this capacity.

Cowan’s Magical Number Four

Some researchers argue that Miller’s seven overestimates true capacity, since his participants were unknowingly chunking items together.

Cowan (2001) reviewed a wide range of memory tasks and found that, once chunking is controlled, capacity is closer to four items.

This lower estimate reframes capacity as a limit of attention rather than a fixed store, a debate still unresolved (see Critical Evaluation).

Duration

Short-term memory typically holds information for only about 15 to 30 seconds. Rehearsal extends it. Repeating the material verbally keeps the trace active for longer.

Atkinson and Shiffrin (1968) built this short duration into their influential multi-store model of memory.

The Brown-Peterson Technique

The classic demonstration of this limit comes from the Brown-Peterson technique, which blocks rehearsal by having participants count backward in threes.

Aim: Peterson and Peterson (1959) tested how long unrehearsed information survives in short-term memory.

Method: Participants saw a trigram of three consonants (e.g., “BRT”). They counted backward in threes from a random number for 3 to 18 seconds. Then they recalled the letters.

Results: Recall fell sharply as the delay grew. About 70% of trigrams were forgotten after nine seconds, and around 90% after 18 seconds.

Conclusion: Even a tiny amount of unrehearsed information is lost within seconds. This is strong evidence that short-term memory has a brief, limited duration.

Why Does STM Forget?

Three explanations compete. The trace may simply decay, fading on its own once time passes. Because capacity is fixed, new items may displace older ones.

Or the counting task itself may interfere, competing with the trigram for the same limited store. The Brown-Peterson task cannot fully separate these three causes, so the true source of forgetting remains debated (see Critical Evaluation).

Short-term memory duration varies between individuals, shaped by attention, distraction, and the type of information involved.

Encoding

Encoding in short-term memory primarily involves a transient representation of information, based mainly on the sensory attributes of the input. Here’s a breakdown of how encoding works:

1. Acoustic Encoding: The most common form of encoding in short-term memory. Verbal information is stored by its sound, which is why you might repeat a phone number aloud to remember it.
2. Visual Encoding: The brief storage of visual images, such as glancing at a picture and recalling its details a moment later.
3. Semantic Encoding: Processing information by its meaning. This dominates long-term memory, though it can help some short-term tasks, like recalling a coherent sentence rather than random words.
4. Tactile Encoding: Storage based on touch, which is less common than acoustic or visual encoding in short-term memory tasks.

Various factors, including attention, repetition, and the nature of the information, can influence the effectiveness of encoding in short-term memory.

Conrad (1964): Acoustic Confusions

Aim: Conrad (1964) tested whether short-term memory codes visual information by sound.

Method: Participants saw sequences of consonants, presented visually one at a time. Each trial used a fresh random order. They had to write down each sequence immediately, in the correct order.

Results: Recall errors were overwhelmingly acoustic confusions rather than visual ones.

Participants often mistook letters like B and P, or V and F, because they sound alike, not because they look alike. Every letter had been seen, not heard.

Conclusion: Short-term memory recodes visual input into a sound-based form. This is why the confusions follow how letters sound rather than how they look on the page. It confirms that acoustic coding dominates even when material never passes through the ears at all.

Baddeley (1966): Acoustic Versus Semantic Coding

Aim: Baddeley (1966) tested whether short-term and long-term memory rely on the same code.

Method: Participants learned one of several word lists, either acoustically similar (e.g., cat, cab, can) or semantically similar (e.g., big, huge, large). Each list was recalled immediately or after a delay.

Results: Immediate recall, which taps short-term memory, was worst for the acoustically similar list.

The sound-alike words were easily confused with each other. Delayed recall, which taps long-term memory, was worst for the semantically similar list instead.

Conclusion: This double dissociation, where each store is disrupted by a different kind of similarity, is strong evidence that short-term and long-term memory are functionally separate systems. They rely on different codes. Neither store can simply be a copy of the other.

However, without further processing, the data held in short-term memory can decay or be displaced, emphasizing the transient nature of this memory store.

More durable and elaborate encoding methods, such as deep processing or the formation of associations, are needed to move information from short-term to long-term memory.

Evidence Short-Term Memory Is a Distinct Store

Two lines of evidence show short-term memory works separately from long-term memory. One comes from how people recall lists of words. The other comes from rare neuropsychological patients whose brain damage affected one memory system but spared the other.

The Serial-Position Effect

When people freely recall a list of words, they remember the first and last items best. Murdock (1962) called this the serial-position effect: a primacy effect for early items and a recency effect for the most recent ones.

Researchers think the two effects have different causes. Early words benefit from extra rehearsal. This transfers them into long-term memory. The most recent words are still sitting in short-term memory when recall begins.

Aim: Glanzer and Cunitz (1966) tested whether the recency effect specifically reflects short-term memory.

Method: They had participants recall word lists either immediately or after a 30-second delay filled with a counting task. The task emptied short-term memory.

Results: Delayed recall abolished the recency effect completely, while the primacy effect stayed intact.

Conclusion: Because only the short-term-based portion of the curve disappeared, this selective loss is strong evidence for a separate, short-lived memory store.

Case Studies: KF and HM

Rare, selective brain damage can affect short-term and long-term memory independently. Patient KF (Shallice & Warrington, 1970) had a severely impaired short-term memory after a brain injury. His digit span, the longest item sequence he could repeat back correctly, was just two.

Most adults manage seven.

Yet his long-term learning and comprehension stayed largely intact. This is the reverse of the usual amnesia pattern. KF’s deficit was also selective: it hit verbal and acoustic memory but largely spared his visual short-term memory.

A different patient shows the opposite pattern.

Patient HM (Scoville & Milner, 1957) showed this after surgery removed tissue on both sides of his brain to treat severe epilepsy. He could still hold information in short-term memory for about 15 seconds. But he could no longer form new long-term memories.

Together, KF and HM form a double dissociation, strong evidence that short-term and long-term memory are separate systems. This split shows up in the brain too. The prefrontal cortex sustains short-term maintenance, while the hippocampus, damaged in HM, supports the transfer to long-term memory (D’Esposito & Postle, 2015).

Working memory

Short-term memory and working memory are closely related but not identical. Short-term memory is passive: it just holds information. Working memory is active: it also manipulates and processes what it holds, acting as the brain’s “workspace” for reasoning and problem-solving.

Baddeley and Hitch (1974) reworked the short-term store into this more active system. Why? A single passive buffer could not carry out the reasoning and comprehension credited to short-term memory.

In place of one store, they proposed a central executive that directs two subsystems.

The phonological loop handles acoustic coding and verbal rehearsal, such as repeating a phone number to keep it active.

It descends from Baddeley and Hitch’s original rehearsal buffer. The visuospatial sketchpad is a separate system that handles visual and spatial information.

Baddeley (2000) later added a third component, the episodic buffer, which integrates the phonological loop and visuospatial sketchpad with long-term memory.

This multi-part view explains findings a single store cannot, like holding a verbal load while reasoning through a different task at the same time.

Applications

Short-term memory’s tight limits on capacity and duration are not just theoretical. They shape how teachers teach, how designers build interfaces, and how clinicians assess brain injury.

Education and Study Skills

Because short-term memory holds so little for so short a time, effective teaching works within those limits. Chunking helps. Grouping single items into meaningful units, like a phone number, lets learners hold more without exceeding the four-to-seven-chunk ceiling.

Presenting too much information at once overloads the store, so little of it is retained. This is the practical core of cognitive-load approaches to lesson design.

Durable learning also depends on getting material into long-term memory. Deep, meaningful processing, such as self-testing, elaboration, and linking new material to prior knowledge, works far better than rote repetition (Craik & Lockhart, 1972).

The serial-position curve has a classroom application too. Material presented in the middle of a lesson is remembered least well, so teachers front-load key points and revisit them later.

Interface and Instructional Design

Designers exploit the same limits. Chunking information, such as spacing digits in a card number or grouping menu options into small sets, keeps the number of items a user must hold in mind low.

External memory aids help too. Visible prompts, checklists, and progress indicators mean users do not need to hold intermediate results in their heads. That is the whole idea.

A banking app that shows a running list of completed steps, rather than asking the user to remember them, applies the same principle.

Verification codes and one-time passwords are typically six digits or fewer for the same reason: a longer unrehearsed string exceeds reliable span and duration.

The same logic favors keeping on-screen instructions visible rather than asking a user to remember a step across screens.

Cognitive Assessment

Short-term memory span is a routine part of clinical assessment. The digit-span task, given forwards and backwards, indexes verbal short-term and working-memory capacity.

It is embedded in standard intelligence and memory batteries used across clinical and educational settings.

An abnormally short span can point to damage in verbal short-term storage, as in patient KF. The reverse pattern is just as telling. A spared span with impaired new learning, as in patient HM, points instead to a problem with transfer into long-term memory.

Because span is sensitive to attention and to the type of material used, clinicians interpret it alongside other measures rather than on its own.

Even so, a single digit-span test can flag when a fuller neuropsychological workup is worth pursuing.

Critical Evaluation

Short-term memory is one of psychology’s most tested ideas, backed by robust classic findings. It also faces real, unresolved questions about exactly what it is and how it works.

Strengths

• Robust, replicated findings: The 7±2 capacity limit (Miller, 1956) and the 15-30 second duration (Peterson & Peterson, 1959) are among the most frequently replicated findings in memory research, giving the model clear boundaries.
• Explains expert performance: Chunking shows how a fixed-capacity store can still handle rich material: grouping digits into a memorable date lets far more pass through the same seven slots, which is why an expert can hold more than a novice.
• Converging evidence: Several independent methods agree: acoustic-based coding, the serial-position effect (only the recency portion collapses under a filled delay), and the KF and HM case studies, where the two systems were damaged separately, showing they can be dissociated both behaviorally and neurologically.
• Practical value: Because the model makes specific, numerical predictions rather than vague claims, it translates directly into practice: chunking-based study techniques, cognitive-load-aware lesson design, and clinical digit-span assessment that helps localize memory impairment all rest on the same capacity-and-duration framework.

Yet several of these classic, well-established findings are now the subject of active ongoing debate:

Limitations

• Not a single, passive store: The working memory model shows short-term memory is multi-component and active, and KF’s selectively verbal deficit shows it is not one uniform system.
• Capacity is disputed: Estimates range from about four chunks (Cowan, 2001) to seven items (Miller, 1956), partly because chunking makes “one item” hard to define.
• Duration findings are confounded: The Brown-Peterson paradigm cannot fully separate memory decay from interference caused by its own counting task, so the true cause of forgetting stays uncertain.
• Separate store debated: Some theorists argue short-term memory is simply activated long-term memory (Cowan, 2008), while others defend it as a genuinely distinct system (Norris, 2017).
• Relies on artificial evidence: Much of the classic evidence comes from meaningless lab tasks and rare, atypical patients, so applying it to everyday memory needs caution.

Contemporary Research

• Attention-based models: Newer accounts, reviewed by Cowan (2008) and D’Esposito and Postle (2015), see short-term storage as temporary activation of long-term memories, controlled by attention rather than a separate box.
• The distinct-store debate continues: Norris (2017) reviewed decades of behavioral evidence and concluded that short-term and long-term memory remain functionally different systems, a position that embedded-processes accounts (Cowan, 2001; Nairne, 2002) still dispute.
• Brain-imaging evidence: D’Esposito and Postle (2015) reviewed neuroimaging and patient studies showing that some short-term content is held in temporary “activity-silent” changes to brain connections, and can be decoded from the same sensory brain regions that first perceived it.

Baddeley (2012) reviews this capacity-versus-attention debate as one of the field’s liveliest open controversies.

References

Atkinson, R. C., & Shiffrin, R. M. (1968). Human memory: A proposed system and its control processes. In K. W. Spence & J. T. Spence (Eds.), The psychology of learning and motivation (Vol. 2, pp. 89–195). Academic Press.

Baddeley, A. D. (1966). The influence of acoustic and semantic similarity on long-term memory for word sequences. Quarterly Journal of Experimental Psychology, 18(4), 302–309.

Baddeley, A. D. (2000). The episodic buffer: A new component of working memory? Trends in Cognitive Sciences, 4(11), 417–423.

Baddeley, A. D. (2012). Working memory: Theories, models, and controversies. Annual Review of Psychology, 63, 1–29.

Baddeley, A. D., & Hitch, G. (1974). Working memory. In G. H. Bower (Ed.), The psychology of learning and motivation (Vol. 8, pp. 47–89). Academic Press.

Conrad, R. (1964). Acoustic confusions in immediate memory. British Journal of Psychology, 55(1), 75–84.

Cowan, N. (2001). The magical number 4 in short-term memory: A reconsideration of mental storage capacity. Behavioral and Brain Sciences, 24(1), 87–114.

Cowan, N. (2008). What are the differences between long-term, short-term, and working memory? Progress in Brain Research, 169, 323–338.

D’Esposito, M., & Postle, B. R. (2015). The cognitive neuroscience of working memory. Annual Review of Psychology, 66, 115–142.

Glanzer, M., & Cunitz, A. R. (1966). Two storage mechanisms in free recall. Journal of Verbal Learning and Verbal Behavior, 5(4), 351–360.

Jacobs, J. (1887). Experiments on “prehension”. Mind, os-12(45), 75–79.

Miller, G. A. (1956). The magical number seven, plus or minus two: Some limits on our capacity for processing information. Psychological Review, 63(2), 81–97.

Murdock, B. B. (1962). The serial position effect of free recall. Journal of Experimental Psychology, 64(5), 482–488.

Nairne, J. S. (2002). Remembering over the short term: The case against the standard model. Annual Review of Psychology, 53, 53–81.

Norris, D. (2017). Short-term memory and long-term memory are still different. Psychological Bulletin, 143(9), 992–1009.

Peterson, L. R., & Peterson, M. J. (1959). Short-term retention of individual verbal items. Journal of Experimental Psychology, 58(3), 193–198.

Scoville, W. B., & Milner, B. (1957). Loss of recent memory after bilateral hippocampal lesions. Journal of Neurology, Neurosurgery, and Psychiatry, 20(1), 11–21.

Shallice, T., & Warrington, E. K. (1970). Independent functioning of verbal memory stores: A neuropsychological study. Quarterly Journal of Experimental Psychology, 22(2), 261–273.