Methodology: how the Dyscalculia.io test is scored

Method last changed October 6, 2026

The free test is 5 short games with 21 fixed items. Each game gives a score from 0 to 1; the result names the highest and the lowest of the 5 and places three skill areas in four bands. There are no age norms and no percentiles, and the result is a screening, not a diagnosis. The items and scoring have not changed since launch on September 29, 2026; every change and correction is dated on the changelog.

What the test measures, and what it does not

The test looks at 5 basic number skills that research links to dyscalculia, one game each. It does not use school maths, so a four-year-old and an adult take exactly the same items.

It does not measure calculation, reading, attention, intelligence or anything about a person beyond these answers on one day. It does not identify dyscalculia and it does not rule it out. Choosing an age group (child, teenager, adult) does not change the items, the scoring or the cut-offs; it only chooses where the practice program starts.

Count the Dots: subitizing, seeing small amounts at a glance.
Which is Bigger?: linking a written number to the amount it stands for.
Which Group Has More?: approximate number sense, judging which group is larger without counting.
Place the Number: the mental number line, placing a number between 0 and 100.
Remember the Digits: working memory for digits.^1,2,3,4,5

The five games and their items

All 21 items are shown in the same order to everyone, so a result can be compared with a later one. There is no time limit on any answer. The time each game takes is recorded, but it does not enter any score.

Items and scoring of each game, in the order they are played
Game	Items	Score of the game
Count the Dots	5 pictures with 3, 6, 2, 8, and 4 dots; the answer is chosen from buttons 1 to 9	Correct answers ÷ 5
Which is Bigger?	5 pairs: 4 vs 9, 7 vs 8, 12 vs 7, 45 vs 53, 67 vs 31; tap the larger number	Correct answers ÷ 5
Which Group Has More?	5 pairs of dot groups: 3 vs 7, 8 vs 9, 12 vs 7, 15 vs 23, 18 vs 22 dots; tap the larger group	Correct answers ÷ 5
Place the Number	3 targets: 25, 67, and 41 on a 0–100 slider that starts at 50	Each item scores 1 − (distance from the target ÷ 100); the game score is the average
Remember the Digits	3 sequences of 3, 4, and 5 digits, each digit shown for 1.4 s with a 0.4 s gap	Sequences repeated in exactly the right order ÷ 3

How the result is built from the five scores

The result shows "Your strength" (the game with the highest score) and "Let's start here" (the game with the lowest score). When two games tie, the one played first is named. Below them, three of the four skill areas get a band; Arithmetic Flow is shown as "Not measured by this test", because none of the games asks for a calculation.

Skill areas and the games behind them
Skill area	Calculated from
Number Sense	Average of Count the Dots and Which Group Has More?
Space & Position	Average of Place the Number and Which is Bigger?
Fact Memory	Remember the Digits
Arithmetic Flow	Not measured by this test

Strong: 0.80 or higher.
Growing: from 0.60 to below 0.80.
Building: from 0.40 to below 0.60.
Focus area: below 0.40.
If the average of all 5 games is below 0.60, the result says that several answers suggest numbers take extra effort and suggests a teacher or an educational psychologist if the difficulty is persistent; otherwise it says most answers came easily.

Where the cut-offs come from

The four bands are equal steps of 0.20 fixed in the code of the Dyscalculia: Numeracy & Math app, and the web test uses the same values so that a result means the same on both. They are the same for every age. They were not derived from a norm sample, a percentile table or a published clinical cut-off, and the 0.60 line for the overall message is simply the lower edge of the "Growing" band.

We have no study that compares this result with a professional assessment, so we make no claim about its accuracy, sensitivity or specificity. The choice of skills follows the research cited on this page; the cut-offs do not come from that research.

What a result means, and what it does not

With 3 to 5 items per game, one wrong answer moves a game score by 0.20 to 0.33, which can be a whole band. A tired moment, a misread instruction or a slip on the slider is enough. A single result should be read as a pattern to check, not as a measurement to rely on.

It means: which of the five skills took the most effort that day, compared with the other four.
It does not mean: a diagnosis, a comparison with other children of the same age, or a level of school maths.
A focus area that matches difficulties seen for months is worth acting on and worth discussing with a teacher or a qualified assessor; a surprising result is worth repeating on another day.
A result in the higher bands does not rule dyscalculia out. Persistent difficulties are a reason for a professional assessment whatever the result says; see how dyscalculia is assessed.

Where the facts on this site come from

Statements about dyscalculia itself (how common it is, its signs, how long it lasts) are linked to a source listed under the section they support. The reference list has 93 entries: 37 studies with a PubMed or DOI record, and 56 other sources such as official statistics, laws, WHO classifications and the published price pages behind the assessment cost tables.

Figures on the dyscalculia statistics page show the year of their own data, not the year we published them. The same figures can be downloaded as CSV and JSON. The test's items, scoring and cut-offs are published as screening-test.json.

What the practice program records

The paid program is not part of the test and never changes a test result. It records practice: each level has 4 questions in the first level of a world and 8 in the others, and gives 3 stars at 90% accuracy, 2 at 75% and 1 at 60%. One star opens the next level.

Within a level, difficulty moves after every block of 2 questions, based on the last 4 answers at the current difficulty: one step up when all were correct and quick, one step down at 50% or less, never more than one step from the level's own difficulty. A level finished at a lower difficulty earns at most 2 stars.

Spaced review: missed questions come back the same day; dot patterns, ten-frames and small sums answered correctly come back after 1, 3, 7, and 30 days. A review session holds up to 8 questions.

If a skill's average over its last 3 sessions drops by 0.15 or more against the sessions before, or stays below 0.60, the daily plan adds a refresh step for it and its review questions come back sooner.

Stars, accuracy and streaks in the parent area describe practice in the program. They are not a re-test and not a measure of dyscalculia.

Changes and corrections

Every change to an item, a score, a cut-off, a program rule, a price or a published figure, and every correction to something the site said, is dated on the changelog, also published as changelog.json. If you find an error, write to [email protected] with the page and what you found.