Every test on this site follows the same rules for working out and showing a result. Each test page adds the details for that test under “How the result is worked out” and “What changes the result”.
What a result shows
- What was measured comes first: the number, its unit and how many tries it is based on, such as the median of 5 tries in milliseconds.
- Then how it was worked out and what changes it. A reaction time, for example, includes the delay of your screen and of your mouse or keyboard, and the page says so.
- Reference values only with a source. When a page shows values from other people, it names the study, its year and who took part. Without such a source, a page shows only your own result.
- Positions only within a study. Where a published study measured the same thing in the same way, a page may say roughly where your result would sit in that study’s sample. It never claims a position among everyone, and grade C tests give none.
- In-between answers stay in between. When a questionnaire sorts answers into types and yours fall near a boundary, the result says so instead of picking one.
- No invented totals. Parts that a published questionnaire scores separately are shown separately here too.
Evidence grades
Each test page shows one of these grades. Puzzles and games have none.
| Grade | What it means | Reference values |
|---|---|---|
| A | The result can be checked against an accepted standard: a published scale that may be used freely, a standard experimental method, or a direct reading of your device. The page names the scale or method and its source. | The scale’s own score bands, or values from a named source |
| B | The method is established, but your device, your surroundings or the questions change the result, so it is a guide. The page says what changes it. | Only when the conditions were similar, and the page says how they differ |
| C | For fun and self-reflection. It has not been validated, and the page says so near the top. | None |
Timing in a browser
Tests that time you use your browser’s clock, which counts in milliseconds. Several things add to a measured time:
- The screen. A 60 Hz screen draws a new picture every 16.7 ms, so a change can appear up to that much later than the page asked for it. Screens with a higher refresh rate shorten the wait.
- Keyboards, mice and touchscreens. Each adds its own delay before the browser learns about a press, and the delay differs from one device to the next. Wireless devices often add more.
- Sound output. Bluetooth headphones and speakers can delay sound by a tenth of a second or more, which matters in tests that play a sound and time your answer.
- The browser’s clock. Browsers round the time that pages can read on purpose, to make some attacks harder. In most browsers the steps are a millisecond or smaller, and privacy settings can make them coarser.
- A busy device. Other tabs, power saving and a slow processor can delay the page.
So a timing result describes this device as well as you. To follow your own progress, take the test on the same device, in the same browser and with the same headphones each time.
Where the questions come from
Quizzes, personality tests and self-checks use one of two kinds of questions:
- A published scale or question bank whose licence allows its use on a website. The test page names it and its authors under References, and the licences page lists its licence.
- Questions written for this site. The page says so. Unless it names a study that checked them, they haven’t been validated.
Health topics
Tests about hearing and sight, and the wellbeing self-checks, are for your own information. Their pages name their sources, say how the result is worked out and when the page was last updated, and say plainly that a result can’t replace a check by a professional. Pages on health topics take a quieter tone: no celebrations, and numbers that appear without animation.
Questions? Email hello@testsonyourdevice.com.
Last updated 13 November 2026