STEM Toys That Actually Teach Something

Buying guide · STEM · Parent-focused

"STEM" is now a sticker. It goes on the box because it moves the box, and it has been applied to things that are, on inspection, a plastic animal in a bag. That's not a scandal — it's just marketing — but it does mean the label has stopped carrying information, and a parent standing in front of a shelf has no way to sort the real ones from the stickered ones.

So here's the sort we use, and it has nothing to do with the word on the box.

The one question that separates them

Can she get a result you didn't plan for?

That's it. A toy where every outcome is in the instruction booklet is a craft activity with a lab coat on. The child follows twelve steps, the volcano erupts, everyone claps, and the thing that was practised was following twelve steps. Useful! Not science.

A real one has a space in it where she can try something nobody suggested and find out what happens. The volcano kit becomes science the moment she asks what happens with twice the vinegar and you say "let's find out" instead of "that's not in the instructions."

Three more tells

It fails visibly. If a thing can only work, there's nothing to diagnose, and diagnosis is most of what science actually is. Good STEM toys break in ways she can see and reason about.

The interesting part isn't the electronics. A lot of "STEM" is a normal toy with a microchip that does the interesting bit on the child's behalf. Watch for what's happening in her hands versus in the circuit board.

It survives being wrong. Materials that get consumed by a failed attempt quietly teach children not to attempt. The best ones let her be wrong ten times for free.

How to play alongside without turning it into a lesson

The fastest way to kill a STEM toy is to start teaching. The moment it becomes a lesson, she's performing for you rather than investigating, and performance and investigation look similar for about four minutes before one of them stops.

What works instead is asking questions you don't know the answer to — and meaning it. What do you think happens if… is a good one. That's odd, why did that one fall over and not that one? is better. Best of all is being visibly wrong yourself: guess out loud, be incorrect, and let her catch you. A child who has watched her dad be wrong about something and find it interesting rather than embarrassing has learned the single most important thing on this page, and no toy is required for it.

Each pick below has a "what to ask" line built into it. Use one, not all of them, and not every time.

Reading the list by age

If you'd rather work from her age than from the concept, the rough ladder is this. Three to five is structures and forces — the ramp and the tiles, where the result is immediate and visible. Five to seven adds materials: the chemistry kit and the wall-climbing car, where she can start changing a variable on purpose. Six to nine is design and iteration, which is where the 3D pen and the print-in-place dragon do their work. Eight and up is the demanding end — the string loop and the aeroplane, both of which reward patience and punish rushing.

For the full age-banded versions of the same catalogue, there's 3–5, 6–8 and 9–12.

What we'd skip

Anything where the "experiment" has one answer. Covered above, but it's the most common failure and worth repeating. If the booklet knows how it ends, it's a recipe.

Coding toys that are mostly a screen. If the physical object is a shell around an app she could have downloaded, you've paid for plastic.

Kits with consumable materials and no refills. A brilliant activity that can be done four times is four activities. Check that it can be topped up before you buy it.

Microscopes and telescopes at the cheapest end of the range. The cheap ones show a blurry grey circle, and a child who has looked into one blurry grey circle has learned that science is boring. If it's not in budget to get a decent one, get something else this year.

The part that isn't a purchase

None of this works if the questions stop. The toy creates the moment; what happens with the moment is entirely down to whether there's an adult nearby who finds the answer genuinely interesting. You do not need to know any physics for this — you need to be willing to say "I don't know, how would we find out?" and then actually find out, badly, together, on a Saturday.

That's the whole method. Everything above is just an excuse to run it.

← Back to all Toy Picks