Digit Span Test — Free Number Memory Span Check
- Free
- ~3 min
- No sign-up
- Instant result
A digit span test: watch digits appear one at a time, then enter them in order. About 3 minutes, climbing from 3 up to 10 digits to find your memory span, with every retake saved so you can track your trend. A free short-term memory self-check.
What's included
- Time
- ~3 min
- Price
- Free
How this digit span test runs
Digits are drawn at random from 0 to 9, and the same digit never appears twice in a row. Each digit stays on screen for 0.8 seconds, followed by a 0.4-second blank, so one digit passes every 1.2 seconds. Then a number pad appears; your answer is scored the moment you enter the last digit, so fix any slip with Delete before that. You start at 3 digits and gain 1 digit per success; after a miss you get one more try at the same length with a fresh sequence. The test ends after 2 misses at the same length, or once you get 10 digits right.
The top of the result screen shows the longest sequence you recalled correctly and your total number of attempts. The detailed report adds your correct and wrong counts, accuracy, your span as a share of the 10-digit maximum, and one line per attempt showing its length and whether you got it right. That lets you tell a clean 7 apart from a 7 reached after one miss. The short interpretation uses three bands: 9 digits or more, 5–8, and 4 or fewer. Because the test stops at 10 digits, if you clear 10 your real span may be longer still.
Beyond 7±2: time and chunks
In the working memory model of Baddeley and Hitch (1974), digits are held in the ‘phonological loop’, which stores sound-based information briefly and keeps it alive through silent rehearsal. Baddeley, Thomson and Buchanan (1975) showed that shorter words are recalled better and that span can be predicted from how many words a person can read in about 2 seconds — so span is tied to rehearsal time, not only to the number of items. Cowan (2001), reviewing studies in which rehearsal and chunking were prevented, concluded that people hold about 3 to 5 chunks at once, around 4 on average.
So the familiar span of around 7 digits is better seen as a short-lived store plus silent rehearsal and grouping, rather than one pure capacity. This test uses forward recall only: you enter the digits in the order shown. A review of order recall notes that with verbal material, performance is usually worse when the sequence must be recalled backward (Donolato et al., 2017); reversing the order adds a manipulation step, which moves the task from holding information toward working with it. If you want to try backward recall too, it is part of the working memory section in Cognitive ability (IQ).
Why language and format change your span
Even when digits are shown on screen, most people silently read them and hold them as sounds. Conrad (1964) found that when visually presented letters were recalled, errors systematically involved letters that sound alike. That is why the time it takes to say digit names matters. Ellis and Hennelly (1980) reported that Welsh–English bilinguals had shorter digit spans in Welsh, whose digit names are longer, and Stigler, Lee and Stevenson (1986) found that Chinese-speaking children, whose digit names are pronounced quickly, remembered more digits than English-speaking children.
Compare your results with your own earlier runs in the same language, not with people who rehearse digits in another. The format differs too: in the Wechsler digit span subtest used in clinics, an examiner reads the digits aloud at 1 per second and you answer out loud, whereas here one digit appears on screen every 1.2 seconds and you type. Hearing a list tends to boost recall of its last items — the modality effect (Crowder and Morton, 1969) — so the two scores are not interchangeable. Use this score as a yardstick against your own records taken the same way, not as a stand-in for a clinical score.
FAQ
What does the digit span test measure?
It measures how much you can hold in mind for a few seconds — your short-term memory span. Digits appear one at a time (about 0.8 seconds each) and you enter them in the same order; while you hold them, your prefrontal and parietal cortices cooperate and the phonological loop silently rehearses the numbers. Devised by Jacobs in 1887, it is one of the longest-serving memory paradigms in psychology.
How many digits counts as a good span?
As psychologist George Miller's famous 'magical number seven, plus or minus two' suggests, most adults top out around 5–9 digits. The test starts at 3 digits, adds one per success, and ends after two misses at the same length; your result is the longest span reached (up to 10). A low score does not indicate any disorder — sleep loss or noise alone can shrink your span. This is a self-check for reference; if worrying symptoms persist, consult a professional about the symptoms themselves, not this score.
Can practice increase my digit span?
Your score on the test itself rises quite a bit with practice — especially once you use chunking, grouping digits in twos and threes like a phone number, which packs a longer sequence into the same capacity. But that reflects better strategy rather than a bigger underlying capacity, and evidence that gains transfer to other memory tasks is limited. Treat it less as a training tool and more as a thermometer for today's memory condition.
How often should I retest, and how do I read the trend?
Every one to two weeks is a good rhythm. Retesting too often mixes in a practice effect from chunking strategies, masking real changes. Results are saved and compared with your previous records on each retake, so watch the trend of spans measured at a similar time of day in a similar condition. If stretches when you keep losing phone numbers or spoken instructions line up with dips in your span, that trend becomes a useful condition signal.
Source
This task is based on the paradigm from: Jacobs, J. (1887). Experiments on prehension. Mind, 12(45), 75–79. This task is a Soundary implementation based on a published academic paradigm. It is not any commercial test product or its test form.
Read more
A Cognitive Psychology Lab in 30 Minutes: Replicating Classic Experiments with Reaction Time, Stroop, Digit Span, Mental Rotation, Trail Making, Tower Planning, and N-back
A lab plan: pick two or three tasks per session, run them with the whole class, and confirm the classic effects in the anonymous distribution. Stroop (1935) and MacLeod (1991); Miller's 7±2 (1956) and Cowan's 4 (2001); Shepard & Metzler's angle–reaction-time line (1971); Reitan (1958) and Tombaugh's (2004) Trail Making norms; Shallice's Tower of London (1982); and the N-back debate from Kirchner (1958) to Jaeggi (2008) and Redick & Lindsey (2013).
11 min read
Teaching Recognition Memory with a Word Memory Test: A 45-Minute Class Plan
A class plan where everyone takes the same five-minute word memory test, the anonymous distribution goes on screen, and hits and false alarms open up recognition memory and signal detection. Classroom-ready numbers from Shepard (1967), Standing's 10,000 pictures (1973), Ebbinghaus's forgetting curve, and Roediger & Karpicke's testing effect (2006).
8 min read
What Is Working Memory, and Does N-Back Training Actually Work?
Working memory is the ability to hold information briefly and work with it. This guide covers the Baddeley model, capacity limits, how digit span, n-back, and spatial span differ, what meta-analyses say about whether n-back training raises intelligence, and how to track your own trend.
5 min read
Related tests
This is a self-screening tool, not a medical diagnosis. If the result concerns you, please consult a mental health professional.
