Skip to content

Digit Span: What the Number Actually Means

A digit span is the longest run of numbers you can repeat back. How the stepped procedure produces that figure, and why an app's best span is not a test score.

By Published Updated 5 min read

Reign Creative LLC makes IQ Test & Brain Training Games, which this article links to, so we have a commercial interest in it. How we write and check articles.

Key takeaways

  • A digit span is the longest sequence of digits you can reproduce in order after a single presentation, and it is produced by a stepped procedure rather than by one attempt at one length.
  • In the standard clinical form used by Choi and colleagues in their normative study, sequences are read at one digit per second, two trials are given at each length, and testing halts when both trials at a length fail.
  • The figure a span task returns is larger than the underlying storage limit, because rehearsal and grouping sit on top of it. Nelson Cowan argues the central store holds roughly 3 to 5 meaningful items in young adults.
  • Norris, Hall and Gathercole found that digit span training does not substantially improve the capacity of verbal short-term memory, and that transfer to other span tasks was minimal. Practice moves the score through better strategy, not a bigger store.
  • IQ Test & Brain Training Games records a best span on its own memory drills, and that is a personal best in a game. Its memory test and IQ-style test are separate features, and its listing describes their results as estimates based on app activities, not clinical or professional assessments.

A digit span is the longest run of digits you can take in once and repeat back in the right order. It is not a number you either have or lack, and it is not read off a single attempt: it is the output of a stepped procedure that lengthens the sequence until you fail. IQ Test & Brain Training Games, which we build at Reign Creative, stores a best span from its own drills, and that figure is a personal best in a game rather than a test score. The app also has a memory test and an IQ-style test, whose results its Google Play listing describes as estimates based on app activities, not clinical or professional assessments; this article explains what the real procedure does so you can see the difference.

What a span task is actually sampling

The task asks you to hold an arbitrary, meaningless sequence in an accessible state for a few seconds and then output it in order. Two things have to go right. You have to keep the items available, and you have to keep their order intact, which is a separate problem — people who recall every digit in the wrong sequence have failed the task just as completely as people who lose two of them.

Because the digits are arbitrary, the task deliberately strips away the thing that normally rescues memory: meaning. You cannot lean on knowledge you already have, which is exactly why the procedure is used to probe a capacity rather than a body of learning. Our guide to working memory and whether training changes it covers the capacity itself in detail; this article is about the measuring procedure that sits on top of it, which the cornerstone does not describe.

How the procedure produces one number

The clinical form is unglamorous and precise. In their normative study of an educationally diverse elderly sample, Choi and colleagues describe using a standardised clinical digit span procedure, presenting digits at a rate of one per second. Forward recall begins at a length of three, backward recall at a length of two. Two trials are presented at each length, and the test is halted when the participant fails both trials at the same digit length.

Three features of that description carry all the weight:

  • A fixed presentation rate. One digit per second is slow enough to be heard and fast enough to block elaborate encoding. Change the rate and you change the score.
  • Two trials per length. A single lucky or unlucky run does not decide anything. Redundancy at each rung is what converts a noisy performance into a stable estimate.
  • A stopping rule agreed in advance. The examiner does not decide when you have had enough. The rule does.

None of those three exists in a casual self-administered version. That is the practical gap between a procedure and a game.

Why the span number overstates your storage

Here the honest answer is uncomfortable for anyone who likes the figure seven. George Miller's 1956 paper observed that decimal digits are worth about 3.3 bits apiece and that we can recall about seven of them. That is where the popular number comes from, and Miller was explicit that the span is a fixed number of chunks rather than of raw items — "we can increase the number of bits of information that it contains simply by building larger and larger chunks."

Nelson Cowan's later review argues that beneath the task surface there is a smaller and more constant central store, limited to roughly 3 to 5 meaningful items in young adults. Standard span tasks come out higher than that because you are allowed to rehearse: Cowan notes that one can retain about two seconds' worth of speech through silent rehearsal, and that in running-span procedures, where rehearsal is disrupted, only the last 3 to 5 digits can be recalled.

So a span of seven is not seven units sitting in a store. It is a smaller core plus rehearsal plus whatever grouping you managed in the moment. That is also why "average digit span" is a slippery phrase: it depends on age, education, the length of the digit words in the language being used, and how much rehearsal the procedure permits. Choi and colleagues published norms for one specific population using one specific instrument, which is the only way a norm can honestly exist.

What practice on a span task does and does not change

This is the question most readers actually arrived with, and the evidence points one way. Norris, Hall and Gathercole, in Memory & Cognition, tested extensive digit span training directly and concluded that digit span training does not substantially improve the capacity of verbal short-term memory, finding little evidence for fundamental capacity changes with extensive practice. Transfer was thin: training on digit span produced no substantial benefit for spoken digits or letter span.

What does improve is your performance on the drilled task, through better grouping, better ordering habits and less flailing under time pressure. That is a real gain and it is worth having. It is not evidence that your memory in general got bigger, and we do not claim otherwise. If you want the full argument about near and far transfer, our evidence review of brain training apps lays out where competent researchers still disagree.

There is a second, quieter reason to be cautious about comparing span numbers across settings. A 2026 study in iScience by Türüdü and colleagues, which followed the same two-trials-per-length structure, reported that a gamified online version produced significantly higher scores than both onsite and standard online versions — and that the boost was identical for forward and backward tasks. A friendlier interface can lift the number without anything about the person changing.

Where the app's drills fit

IQ Test & Brain Training Games ships memory modes built on exactly this shape: a Digit Span Trainer that runs twelve rounds and lengthens the digit string as you go, plus Number Recall and Sequence Memory, which run up to twenty items at the higher difficulty levels. The app keeps your longest completed sequence as a personal best, alongside score, accuracy, streak and response time.

Treat that record as what it is. It was produced on your own phone, at a moment of your choosing, with no examiner and no standard presentation rate you did not set. It tells you whether this week's practice went better than last week's on the same mode at the same difficulty. It does not tell you anything about your standing relative to other people. Keeping practice, entertainment and standardised psychological assessment in separate boxes is not a legal formality — it is the only way the practice record stays useful.

If you want to use the drills seriously, hold the mode and the difficulty constant, and read accuracy before you read the span. You can see how the app describes its memory games and memory test on the IQ Test & Brain Training Games app page, browse the rest of this subject in the memory and attention hub, and find the wider set of study and cognition articles in the education and brain section. For the strategies that actually move a span — grouping, rhythm and rehearsal — the practical companion piece is our comparison of number memorisation methods.

Good to know

Frequently asked questions

What is a digit span?
It is the longest sequence of digits a person can repeat back in the correct order after hearing or seeing it once. The span is not read off a single attempt. Sequences start short, grow by one digit at a time, and the procedure stops once you fail the trials at a given length. The number reported is the last length you handled reliably.
How is the digit span task administered?
Choi and colleagues, using a standardised clinical form of the task in a normative study, describe presenting digits at a speed of one per second, beginning at a length of three for forward recall and two for backward recall. Two trials are given at each length, and testing is halted when the participant fails both trials at the same length.
How many digits can most people remember?
The figure people usually quote traces to George Miller's 1956 paper, which noted that we can recall about seven decimal digits. That number is not a fixed human constant. It shifts with age, education, how long the digit words take to say in your language, and whether you are allowed to rehearse. Treat any single figure as a rough landmark rather than a norm.
Can practice increase your digit span?
Your score on the practised task can rise. Norris, Hall and Gathercole concluded that digit span training does not substantially improve the capacity of verbal short-term memory, and found little evidence of fundamental capacity change with extensive practice. Improvement generally reflects better grouping and strategy on that specific task rather than a larger underlying store.
Does IQ Test & Brain Training Games give me a digit span score?
It records the longest sequence you have completed in its own memory drills and keeps that as a personal best — a game statistic produced under game conditions, with no standard administration. The app also has a memory test you can repeat against your own history, and its listing describes that result, like its other test results, as an estimate based on app activities rather than a clinical or professional assessment. Neither is comparable to a score produced by a psychologist using a standardised instrument.

Sources

  1. A Normative Study of the Digit Span in an Educationally Diverse Elderly Population (opens in a new tab)Psychiatry Investigation, via PubMed Central (PMC3942550) — accessed
  2. The Magical Mystery Four: How is Working Memory Capacity Limited, and Why? (opens in a new tab)Current Directions in Psychological Science, via PubMed Central (PMC2864034) — accessed
  3. Can short-term memory be trained? (opens in a new tab)Memory & Cognition, via PubMed Central (PMC6647533) — accessed
  4. The Magical Number Seven, Plus or Minus Two: Some Limits on our Capacity for Processing Information (opens in a new tab)Psychological Review (1956), via Classics in the History of Psychology, York University — accessed
  5. Platform, modality, and gamification effects on digit span task performance (opens in a new tab)iScience, via PubMed Central (PMC13214330) — accessed

About the author

— The independent app studio that builds the apps covered on this site and publishes every article here. On Google Play it publishes as Reign Collective Apps.

We use AI tools to help research, draft and edit our articles. Each article lists the sources its factual claims rely on, and details about our apps come from their live Google Play listings. No article on this site has been reviewed by an independent expert unless the reviewer is named on it.

Spotted an error? Tell us — how corrections are handled is set out in our editorial policy.

Keep reading