A standard for learning apps
The Fifteen-Minute Test
Every question here can be answered by spending a quarter of an hour with an app and watching what it does. You do not need to know anything about software. It is the standard our own apps are held to, and their scores are at the bottom, including the places they fall short.
First, the test
Do these things in this order before deciding anything. Write down what you see. The scoring afterwards is only a way of adding up what you wrote down.
Open it and touch nothing for thirty seconds.What happens by itself? A good app waits for the child. A bad one starts moving, playing sound or showing a countdown before anyone has asked it to.
Try to stop.Look for the end. Is there a point where the app has nothing more to show and says so? Now try to leave in the middle. Does it let you go, or does it ask you to stay, warn you about losing something, or offer one more thing?
Find out what it wants to know about your child.Look at the first screens and the settings. Does it ask for a name, an age, a birthday, a photograph or the microphone? Does it need those to do its job, or does it just want them?
Find a fact and ask how they know.Pick any claim the app makes. Is there an explanation of how anybody found that out? Does the app say when something is uncertain, or does it state everything with the same confidence?
Leave it alone for five minutes.Put it down with the screen on and walk away. When you come back, has it done anything to get your attention? Sounds, animations, a message?
Turn the sound off and try again.Does the app still work? If it relies on sound to reward the child every time they do something, you will notice the silence very quickly.
Give it to your child, and after ten minutes ask them to stop.This is the most important test and the only one that needs the child. A good app lets go. A bad app makes stopping a fight, and it is the app doing that, not the child.
Then, the score
Two questions. Does the app take from the child, or leave them alone? And does it give the child anything real? Six signs for each. Tap the description that matches what you saw.
A. Does it leave the child's attention alone?
B. Does it give the child something real?
Red flags
If you saw any of these, tick it. The score no longer matters.
Reading the score
| Score | What it means |
|---|
| 20–24 | Give it to your child and do not worry about it. |
| 14–19 | Fine. Look at the signs that scored low and decide whether they matter to you. |
| 8–13 | Be careful. The app is doing some work on the child rather than for them. |
| Under 8 | Do not. |
Two things a total can hide. A zero on A2 (rewards) or A6 (letting go) counts as a red flag whatever the total; those are the two signs the research is clearest about. And a high score on A with a low score on B means the app is harmless and empty. That is not nothing, but it is not a learning app either, and the time might be better spent on a book, a walk or a conversation.
The question behind all of it
Before scoring anything, ask: would this be better as a book, a walk or a conversation?
Most learning apps would be. A screen can do a few things a book cannot. It can show true scale, let a child move through something and watch it change, and hold more than a book can hold. If the app does not use one of those, it has to explain why it is on a screen at all.
Our own apps exist because a book cannot show a child what the solar system looks like at true scale, or what a billion years is beside a lifetime. That is their reason for being on a screen. It is fair to ask any app for its reason.
What to ask the people who made it
If you cannot answer a question above by looking, ask the maker. Any of these is fair, and a good app has a short answer.
- Does it have an end?
- What does it store, and where?
- Why is there a reward there?
- What does it do while I am not looking?
- Why does it need to know that about my child?
- How did you check that the facts are right?
A maker who cannot answer, or who answers with a paragraph about engagement, has told you what you needed to know.
How our own apps score
Scored honestly, with the weak signs named, because a standard that only its author passes is not a standard. The apps are How Big?, How Far? and How Long?, one road each through size, distance and time, and The Long Way Round, which is the orbits in three dimensions.
| Sign | Big | Far | Long | Round | Notes |
|---|
| A1 · an end | 2 | 2 | 2 | 2 | Both ends of each road say so. The orrery is an explorer rather than a road and had no arrival at all; it now announces the two ends it has. |
| A2 · rewards | 2 | 2 | 2 | 2 | No points, stars, sounds or celebrations anywhere. The new questions are unscored: the answer is stated, and that is all. |
| A3 · stillness | 2 | 2 | 2 | 2 | Nothing moves unless the child moves it. |
| A4 · quiet when closed | 2 | 2 | 2 | 2 | No notifications, no account, no network. Guesses are held for the session and then forgotten. |
| A5 · asks nothing | 2 | 2 | 2 | 2 | Nothing leaves the device. The orrery's sky readout gives the angle from the Sun, which is the same for everybody, rather than asking where anyone is. |
| A6 · lets go | 2 | 2 | 2 | 2 | Designed to. Still not tested with a child asked to stop. |
| B1 · true | 2 | 2 | 2 | 2 | Numbers checked against published values; estimates say so. The orrery is checked against JPL Horizons on every test run. |
| B2 · explains | 1 | 2 | 2 | 2 | How Big? carries less how we know than the others. |
| B3 · the child thinks | 1 | 2 | 1 | 2 | Fixed in two of the four: every stop opens with an optional, unscored question answerable before the page gives anything away. |
| B4 · reaches outside | 1 | 2 | 1 | 2 | Fixed in the same two. How Far? shrinks the Sun to a football and gives every stop in paces. The orrery works out whether a planet is a morning or an evening object today. |
| B5 · capable person | 2 | 2 | 2 | 2 | Real words, explained. No mascot, no squeak. |
| B6 · honest unknowns | 2 | 2 | 2 | 2 | Every stop ends on something nobody knows. |
| Total | 21 | 24 | 22 | 24 | out of 24 |
How Far? and The Long Way Round were the two taken to 24. What lifted them is exactly what this page said was missing: a question before the reveal, and something to do away from the screen. How Far? shrinks the Sun to a 22 cm football, which makes Earth a 2 mm bead 24 paces away and the nearest star 6,300 km. The Long Way Round computes each planet's angle from the Sun for the date shown and says whether it is up in the evening or the morning.
How Big? and How Long? were deliberately left alone, so their low marks stand and are the next work. The fix is known and it is the same one.
A6 is the exception, and it is the most important sign here. The apps are built to let go and nothing in them warns, offers or delays. But no child has been asked to put one down while somebody watched, so the 2 is an argument rather than an observation, and the honest way to read both 24s is 24 with one sign taken on trust. That test needs a person, ten minutes and a child. It is step 7 above, and it is the only part of this standard its author cannot run alone.
Where this comes from
The six attention signs come from the research on children, screens and attention. It points much more clearly at not being able to stop than at screen time itself, and it finds that persuasive design harms most the children who can least resist it.
The six learning signs come from building these apps and finding out what mattered: that a fact without a how is not learned, that a child treated as capable rises to it, that the unknown is the most interesting part, and that an app is only worth its screen if it does something a book cannot.