How Many Items Can the Average Person Hold in Working Memory?

Miller vs. Cowan: the revised number

In 1956, George Miller published "The Magical Number Seven, Plus or Minus Two," arguing that human short-term memory capacity was approximately 7±2 items. The paper became the most-cited article in psychology history. But Miller's findings were based on tasks that allowed participants to use chunking and rehearsal strategies — meaning the "seven" included all the cognitive tricks people use to extend apparent capacity, not the raw storage limit.

In 2001, Nelson Cowan published a thorough meta-analysis and theoretical synthesis concluding that true working memory capacity — stripped of rehearsal and chunking — is approximately 4 chunks, with an individual range of roughly 3–5. This is now the consensus in cognitive neuroscience. The distinction matters: 7±2 describes performance, while 4 describes fundamental capacity.

The key distinction

Miller's 7±2 (performance)

Includes chunking, rehearsal, and long-term memory assistance. Explains why people can recall 7-digit phone numbers — they chunk "555" as a unit.

Cowan's 4 (raw capacity)

The true number of discrete items held simultaneously. Measured by preventing rehearsal and using novel, unchunkable stimuli. Confirmed by neuroimaging.

What counts as "one item"?

The answer to "how many items" depends entirely on how you define an item — and this is where the magic of working memory lies. A chunk is any unit of information that has been bound together through prior learning or current strategy. The chess grandmaster holds a meaningful board position as a single chunk; a novice holds it as 30 separate pieces. Both have the same raw working memory capacity of ~4 chunks.

Domain	Novice chunk	Expert chunk	Expert advantage
Chess	1 piece	6–8 piece formation	6–8×
Music reading	1 note	4-bar phrase	~16×
Computer code	1 token	Design pattern	10–20×
Sport tactics	1 player position	Full formation	5–11×
Phone numbers	1 digit	3-digit area code	3×

This explains why expertise in a domain effectively expands working memory for that domain — without changing the underlying 4-chunk limit. It also explains why the Number Memory test gets dramatically easier once you learn to chunk digit groups, and why the Sequence Memory test responds to the same chunking strategy.

Individual variation and what drives it

Although the average is 4 chunks, there is meaningful individual variation. Most people fall between 3 and 5 chunks; high-capacity individuals reach 6–7 even with chunking controlled. This variation is largely heritable (estimated at 50% genetic) but is also shaped by age, exercise, sleep quality, and practice.

Distribution of working memory capacity (pure chunks, controlled)

Practical implications of the 4-item limit

Instruction and learning design

Direct application

Presenting more than 4 new concepts simultaneously guarantees working memory overload — regardless of intelligence. This is why effective teaching sequences no more than 3–4 new ideas before consolidating, and why well-designed user interfaces limit options per screen. The 4-item limit is not a weakness to overcome; it is a hardware constraint to design around. Our article on working memory exercises for students and professionals explores applied strategies for this limit.

Meeting and presentation design

Direct application

Presentations that exceed 4 key points per slide guarantee that the audience will not process all of them. The cognitive science of presentations — pioneered by Richard Mayer and later John Medina — recommends limiting slides to 3 main points, using visuals rather than text to reduce phonological loop load, and pausing after every 4th concept to allow encoding into long-term memory.

Task management under load

Indirect application

The 4-item limit explains why written to-do lists are so effective — they offload working memory to external storage. Every item kept in your head "uses a slot" that could otherwise be used for active processing. GTD (Getting Things Done) methodology is, at its core, a system for externalizing the contents of working memory to free up capacity for the task at hand. This connects to why working memory matters in everyday life beyond simple test scores.

Can you genuinely expand the limit?

The honest answer is: the 4-chunk ceiling is difficult to move, but not immovable. The most reliable strategy is increasing chunk size through deliberate practice in a domain — not increasing the number of chunks. However, certain interventions show small but real increases in raw capacity:

What genuinely helps

→ Aerobic exercise (increases BDNF, preserves prefrontal volume)
→ Sustained sleep hygiene (consolidates capacity, prevents decline)
→ Mindfulness training (+0.5 chunk on average across 8 weeks)
→ Domain expertise (expands effective capacity without moving the ceiling)

The full evidence on long-term capacity training is reviewed in our article on whether you can train working memory long-term. To benchmark where you currently stand, compare your Sequence Memory and Number Memory scores.