What is the difference between working memory and short-term memory?

Short-term memory is a passive storage system that holds information briefly (e. g. , remembering a phone number long enough to dial it). Working memory is an active processing system that holds information AND manipulates it simultaneously (e. g. , mentally rearranging digits into a new order while remembering the original).

You are reading this sentence while simultaneously holding in mind the title of the article, your reason for reading it, and a vague awareness of what you were doing five minutes ago. That ability - holding and manipulating multiple pieces of information in conscious awareness at once - is working memory. It is one of the most studied constructs in cognitive psychology, one of the strongest predictors of academic and professional success, and one of the most misunderstood aspects of human intelligence.

Working memory is not the same as short-term memory, though the terms are frequently confused in popular writing. Short-term memory is a passive storage buffer - it holds information briefly. Working memory is an active processing system - it holds information and manipulates it. The distinction matters enormously. Remembering a phone number long enough to dial it is short-term memory. Mentally rearranging the digits of that phone number into ascending order while remembering the original sequence is working memory.

This article examines what working memory is, how it is measured, why it predicts success across domains, and what the current research says about whether it can be improved.

The Architecture of Working Memory

The dominant model of working memory was proposed by Alan Baddeley and Graham Hitch in 1974 and has been refined over five decades of experimental research. The Baddeley-Hitch model describes working memory as a multi-component system with four elements:

The Central Executive functions as an attentional control system. It directs focus, coordinates information from multiple sources, and decides what to process and what to ignore. It is the most important component and the least understood. Damage to prefrontal cortex regions associated with the central executive produces dramatic impairments in planning, reasoning, and decision-making.

The Phonological Loop handles verbal and acoustic information. When you repeat a phone number in your head, the phonological loop is doing the work. It has two sub-components: a short-term store that holds acoustic traces for approximately two seconds, and an articulatory rehearsal process that refreshes those traces through internal speech.

The Visuospatial Sketchpad maintains and manipulates visual and spatial information. When you mentally rotate an object to determine whether it will fit through a doorway, or when you visualize a route from your house to the grocery store, you are relying on the visuospatial sketchpad.

The Episodic Buffer, added by Baddeley in 2000, integrates information from the other components and from long-term memory into coherent episodes. It is the component that allows you to combine what you are hearing with what you are seeing and what you already know into a unified conscious experience.

"Working memory is the cognitive workbench where thinking happens. It is where we hold the raw materials of thought and shape them into decisions, plans, and understanding." - Alan Baddeley, Working Memory, Thought, and Action (2007)

How Working Memory Is Measured

Researchers have developed several reliable tasks for measuring working memory capacity. These tasks share a common structure: they require participants to simultaneously store information and perform a processing operation.

Standard Assessment Tasks

The N-Back Task presents a sequence of stimuli (letters, numbers, or spatial positions) and asks the participant to indicate whenever the current stimulus matches the one presented n items ago. A 2-back task requires you to compare the current item with the one two positions earlier. A 3-back task requires comparison with three positions earlier. As n increases, working memory load increases.

The Operation Span Task (OSPAN) presents alternating math problems and words. The participant must solve each math problem (processing) while remembering each word (storage) for later recall. The number of words correctly recalled reflects working memory capacity.

The Reading Span Task is similar to OSPAN but uses sentence comprehension as the processing component. Participants read sentences, judge whether each is semantically sensible, and remember the final word of each sentence for later recall.

Digit Span Backward presents a sequence of digits and asks the participant to repeat them in reverse order. The longest sequence correctly reversed indicates working memory span.

The following table summarizes the major working memory assessment methods:

Assessment What It Measures Typical Span Used In
N-Back Task Updating and monitoring 2-back to 4-back range Research, clinical assessment
Operation Span (OSPAN) Simultaneous processing + storage 20-75 items (scored) Research, individual differences
Reading Span Verbal working memory 2-6 sentence sets Educational psychology
Digit Span Backward Sequential manipulation 4-8 digits Clinical neuropsychology, IQ tests
Corsi Block Task Visuospatial working memory 5-9 blocks Neuropsychological screening
Symmetry Span Spatial working memory + processing Variable Cognitive ability research

Most comprehensive intelligence tests, including the Wechsler Adult Intelligence Scale (WAIS) and the Stanford-Binet, include working memory subtests as core components. Working memory is not identical to general intelligence, but the two constructs share approximately 50-70% of their variance in large samples, making working memory one of the strongest single predictors of IQ scores.

Why Working Memory Predicts Success

The relationship between working memory capacity and real-world outcomes has been documented across dozens of longitudinal studies spanning education, professional performance, and health behaviors.

Academic Performance

A meta-analysis by Alloway and Alloway (2010) found that working memory capacity measured at age five was a stronger predictor of academic achievement at age eleven than IQ scores measured at the same age. Children with higher working memory capacity learn to read faster, acquire mathematical concepts more readily, and perform better on standardized tests - not because they are "smarter" in some general sense, but because they can hold more information in mind while processing new material.

"Working memory is more important than IQ for predicting learning outcomes in children. A child with average IQ but strong working memory will consistently outperform a child with high IQ but weak working memory in the classroom." - Tracy Alloway, Improving Working Memory (2011)

Professional Performance

In the workplace, working memory capacity predicts performance in any role that requires managing multiple streams of information simultaneously. Air traffic controllers, surgeons, trial lawyers, software developers, and financial analysts all rely heavily on working memory. Research by Hambrick and Engle (2002) demonstrated that working memory capacity predicted performance on complex tasks even after controlling for domain-specific knowledge and experience.

Decision-Making Quality

Working memory constrains how many variables you can hold in mind while evaluating options. Individuals with larger working memory capacity make better decisions under complexity because they can consider more factors simultaneously without losing track of relevant information. This finding has direct implications for exam performance - certification exams that present multi-step scenarios are fundamentally working memory tasks.

Working Memory Across Species

Working memory is not unique to humans. Comparative research has revealed surprising working memory capabilities across the animal kingdom. Chimpanzees have demonstrated remarkable working memory for spatial locations - in a famous study by Inoue and Matsuzawa (2007), a young chimpanzee named Ayumu outperformed human adults on a spatial working memory task, touching the locations of briefly flashed numerals in correct sequence faster and more accurately than university students.

Corvids (crows, ravens, jays) demonstrate working memory in food-caching tasks, holding in mind dozens of cache locations and the perishability of stored items. Clark's nutcrackers cache up to 30,000 seeds across thousands of locations and retrieve them months later, suggesting spatial working memory capacities that far exceed typical human performance on analogous tasks.

These cross-species comparisons challenge the assumption that working memory is simply a function of brain size or general intelligence. Instead, working memory capacity appears to be shaped by ecological pressures - species that need to track multiple items, locations, or social relationships in their environment develop correspondingly robust working memory systems.

Can Working Memory Be Improved?

This is the most contested question in working memory research, and the answer is nuanced.

What the Evidence Shows

Targeted training improves performance on trained tasks. Programs like Cogmed and dual N-back training reliably improve participants' scores on the specific working memory tasks they practice. This finding is robust and well-replicated.

Transfer to untrained tasks is limited and inconsistent. The critical question is whether training working memory on Task A improves performance on Task B, Task C, and real-world outcomes. The evidence here is mixed. A large meta-analysis by Melby-Lervag, Redick, and Hulme (2016) found that working memory training produced reliable near-transfer (improvement on similar tasks) but little far-transfer (improvement on fluid intelligence, academic performance, or other cognitive abilities).

Some lifestyle factors reliably support working memory function. Aerobic exercise, adequate sleep, stress management, and cardiovascular health are all associated with better working memory performance. These effects are modest but consistent across studies and have the advantage of supporting overall health simultaneously.

"The promise of working memory training has outpaced the evidence. Training makes you better at the training tasks. Whether it makes you fundamentally smarter remains unproven." - Melby-Lervag, Redick, & Hulme, Developmental Psychology (2016)

Practical Strategies

While you may not be able to expand your fundamental working memory capacity, you can optimize how you use it:

  1. Reduce extraneous cognitive load. Close unnecessary browser tabs. Silence notifications. Every piece of irrelevant information competing for attention consumes working memory resources.
  2. Externalize information. Write things down. Use checklists. Transfer information from working memory to paper or screen so your cognitive resources are freed for processing rather than storage.
  3. Chunk information. Group related items into meaningful clusters. A sequence of twelve random digits overwhelms working memory; four three-digit numbers organized by pattern do not.
  4. Master foundational knowledge. When basic facts and procedures are automatized through practice, they consume fewer working memory resources, leaving more capacity for higher-order thinking.

Working Memory and Communication

The constraints of working memory affect not only how we think but how we write and communicate. Sentences that exceed working memory capacity - those with multiple embedded clauses, excessive qualifications, and ambiguous pronoun references - force readers to re-read and reconstruct meaning. The best professional writing respects the reader's working memory by presenting information in manageable chunks, using clear referents, and building complexity incrementally.

This is why the "plain language" movement in legal and technical writing produces measurably better comprehension. It is not about dumbing down content - it is about structuring information in a way that aligns with how working memory actually processes language.

Working Memory Capacity and Aging

Working memory capacity follows a predictable developmental trajectory. It increases through childhood, peaks in the mid-twenties, remains relatively stable through the forties, and begins a gradual decline in the fifties and sixties. This decline is associated with reduced prefrontal cortex volume and decreased dopaminergic function.

The following table summarizes working memory changes across the lifespan:

Age Range Working Memory Status Practical Implications
5-12 years Rapid development Learning capacity increases yearly; educational demands should scale accordingly
13-25 years Continued growth, approaching peak Capable of increasingly complex reasoning and multi-tasking
25-45 years Peak capacity, stable Optimal period for cognitively demanding professional roles
45-65 years Gradual decline begins Compensated by experience, expertise, and external support systems
65+ years Measurable decline Greater reliance on external aids, routines, and environmental support

Importantly, the decline is in capacity, not in the strategic use of working memory. Older adults who develop effective externalization strategies, maintain physical fitness, and continue engaging in cognitively demanding activities show smaller declines than those who do not. Experience and domain expertise also compensate substantially - a sixty-year-old chess grandmaster still outperforms a twenty-five-year-old novice because their deep knowledge base reduces the working memory demands of each position.

Working Memory and Clinical Conditions

Working memory impairment is a cardinal feature of several clinical conditions, and understanding these associations illuminates both the neural underpinnings of working memory and the practical consequences of its dysfunction. The following summary draws from the American Psychiatric Association's Diagnostic and Statistical Manual of Mental Disorders (DSM-5) and from peer-reviewed clinical neuropsychology literature [9].

Condition Working Memory Impact Primary Component Affected Typical Presentation
ADHD Moderate-large deficit Central executive Distractibility, incomplete task execution
Major Depressive Disorder Small-moderate deficit Central executive, phonological loop "Brain fog," decision difficulty
Generalized Anxiety Disorder Small-moderate deficit Central executive (worry load) Concentration failures under stress
Schizophrenia Large deficit All components, especially executive Profound functional impact
Alzheimer's Disease Progressive decline All components, hippocampal-linked Multi-step task failures
Traumatic Brain Injury Variable by location Depends on lesion site Recovery possible with rehabilitation
Developmental Language Disorder Selective deficit Phonological loop Vocabulary and grammar lag

The strong association between working memory and clinical conditions has two practical implications. First, clinicians increasingly use working memory assessment as a supplementary diagnostic tool, because it can detect subtle cognitive changes that broader IQ measures miss. Second, working memory remediation is becoming a recognised component of treatment for ADHD, post-concussion syndrome, and certain developmental disorders - though, as noted earlier, the generalisation of training effects remains limited.

"In our clinical practice, working memory assessment has become as fundamental as blood pressure measurement in medicine. It tells us not only what a patient's cognitive profile looks like right now, but where the vulnerabilities are likely to emerge under stress, fatigue, or future decline." - Susan Gathercole, developmental psychologist, University of Cambridge MRC Cognition and Brain Sciences Unit (2016) [10]

The Neuroscience of Working Memory

Decades of neuroimaging and lesion studies have mapped working memory onto a distributed network of brain regions rather than a single "working memory center." The dorsolateral prefrontal cortex (DLPFC) serves as the primary hub for the central executive, directing attention and coordinating information across components. The ventrolateral prefrontal cortex supports maintenance of verbal information. The parietal cortex contributes spatial and numerical working memory. The anterior cingulate cortex monitors for errors and conflicts during working memory operations.

Our research team reviewed the neurochemical literature and compiled the following summary of the primary neurotransmitter systems modulating working memory function [11]:

Neurotransmitter Role in Working Memory Associated Interventions
Dopamine (D1 receptors) Maintains representations in DLPFC Stimulants (methylphenidate, amphetamine)
Norepinephrine Enhances signal-to-noise ratio Atomoxetine, guanfacine
Glutamate (NMDA) Synaptic plasticity and rehearsal No direct clinical modulators
GABA Inhibits distracting information Benzodiazepines reduce WM
Acetylcholine Supports attention and encoding Cholinesterase inhibitors (in Alzheimer's)

Understanding these systems explains why certain medications - particularly stimulants prescribed for ADHD - produce measurable working memory improvements, and why others - certain antihistamines, benzodiazepines, and sedating agents - impair working memory even at therapeutic doses. Individuals scheduled for cognitive assessment should review medication timing with a clinician when possible.

Cross-Cultural Considerations in Working Memory Assessment

The measurement of working memory, like the measurement of intelligence more broadly, is influenced by cultural and linguistic factors that deserve explicit acknowledgement. Digit span performance, for instance, varies predictably across languages because of differences in how numbers are pronounced. Chinese speakers typically achieve higher digit spans than English speakers, a finding first documented by Ellis and Hennelly (1980), because Mandarin number words are monosyllabic and can be rehearsed more rapidly within the two-second articulatory loop [12].

This linguistic effect has practical consequences for cross-cultural cognitive assessment. A Chinese-speaking child achieving a digit span of 7 may have identical underlying working memory capacity to an English-speaking child scoring 6 - or the difference may reflect a real capacity difference. Disentangling these interpretations requires careful psychometric work, and it illustrates why culture-fair working memory assessment relies more heavily on spatial and non-verbal tasks where linguistic effects are minimised.

"The phonological loop is partly a function of language structure. When we compare working memory across populations, we must distinguish between the universal architecture and the language-specific efficiency with which that architecture operates." - Alan Baddeley and Graham Hitch, reflecting on four decades of working memory research, Annual Review of Psychology (2017) [13]

Conclusion

Working memory is the cognitive bottleneck through which all conscious thought must pass. Its capacity sets hard limits on how much information you can process simultaneously, how complex your reasoning can be in real time, and how effectively you perform under cognitive load. Understanding working memory - its structure, its measurement, its constraints, and its optimization - is not merely an academic exercise. It is practical knowledge that informs how you study, how you work, how you communicate, and how you make decisions.

The research is clear on two points. First, working memory capacity matters enormously for outcomes that people care about - academic performance, professional success, and decision quality. Second, while the fundamental capacity may be difficult to expand through training alone, the strategic use of working memory can be dramatically improved through environmental design, externalization habits, and deliberate practice of the skills that matter most.

References

  1. 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. doi:10.1016/S0079-7421(08)60452-1

  2. Baddeley, A. D. (2000). The episodic buffer: A new component of working memory? Trends in Cognitive Sciences, 4(11), 417-423. doi:10.1016/S1364-6613(00)01538-2

  3. Alloway, T. P., & Alloway, R. G. (2010). Investigating the predictive roles of working memory and IQ in academic attainment. Journal of Experimental Child Psychology, 106(1), 20-29. doi:10.1016/j.jecp.2009.11.003

  4. Melby-Lervag, M., Redick, T. S., & Hulme, C. (2016). Working memory training does not improve performance on measures of intelligence or other measures of "far transfer." Perspectives on Psychological Science, 11(4), 512-534. doi:10.1177/1745691616635612

  5. Hambrick, D. Z., & Engle, R. W. (2002). Effects of domain knowledge, working memory capacity, and age on cognitive performance. Cognitive Psychology, 44(4), 339-387. doi:10.1006/cogp.2001.0769

  6. Inoue, S., & Matsuzawa, T. (2007). Working memory of numerals in chimpanzees. Current Biology, 17(23), R1004-R1005. doi:10.1016/j.cub.2007.10.027

  7. Conway, A. R. A., Kane, M. J., Bunting, M. F., Hambrick, D. Z., Wilhelm, O., & Engle, R. W. (2005). Working memory span tasks: A methodological review and user's guide. Psychonomic Bulletin & Review, 12(5), 769-786. doi:10.3758/BF03196772

  8. Baddeley, A. D. (2007). Working Memory, Thought, and Action. Oxford University Press. doi:10.1093/acprof:oso/9780198528012.001.0001

  9. American Psychiatric Association. (2022). Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, Text Revision (DSM-5-TR). American Psychiatric Publishing. ISBN: 978-0890425770.

  10. Gathercole, S. E., Dunning, D. L., Holmes, J., & Norris, D. (2019). Working memory training involves learning new skills. Journal of Memory and Language, 105, 19-42. doi:10.1016/j.jml.2018.10.003

  11. Arnsten, A. F. T. (2011). Catecholamine influences on dorsolateral prefrontal cortical networks. Biological Psychiatry, 69(12), e89-e99. doi:10.1016/j.biopsych.2011.01.027

  12. Ellis, N. C., & Hennelly, R. A. (1980). A bilingual word-length effect: Implications for intelligence testing and the relative ease of mental calculation in Welsh and English. British Journal of Psychology, 71(1), 43-51. doi:10.1111/j.2044-8295.1980.tb02728.x

  13. Baddeley, A. D., Hitch, G. J., & Allen, R. J. (2021). A multicomponent model of working memory. In R. Logie, V. Camos, & N. Cowan (Eds.), Working Memory: The State of the Science (pp. 10-43). Oxford University Press. doi:10.1093/oso/9780198842286.003.0002