The opposite problem is quieter and gets noticed later. A child builds every supplied model beautifully, finishes each one, and then sits with a full box of parts saying there is nothing to build. Adults tend to be pleased with the first half of that sentence and puzzled by the second. It is a real limitation, and it responds to a gentle, staged approach rather than to being told to use their imagination.

Why the blank page is hard

Following a booklet gives a guaranteed outcome, a visible finish line and no risk of being wrong. Free building removes all three at once. For a child who values completion, that is not liberating, it is unpleasant. Telling them to just make something up asks them to give up everything that made the activity enjoyable, so of course they decline.

Modify, do not invent

The bridge from one mode to the other is alteration. Rebuild a supplied model, then change exactly one thing: longer span, wheels on the base, the drive moved to the other end. The finished object still exists, the risk is contained to one decision, and the child has designed something without being asked to face an empty table. Model ranges like K’NEX include several builds per box, which gives you a supply of starting points for this.

Constraints as a substitute for instructions

A specification behaves like a booklet in the ways that matter: it defines success and it ends. Build something that spans the gap between two chairs and holds a book. Build the tallest thing that survives a nudge. There is still a clear pass or fail, which is exactly what this builder wants, but the route is open.

A staged programme over a few weeks

• Week one, build a supplied model exactly as drawn.
• Week two, rebuild it with one deliberate change.
• Week three, build to a written specification with no picture.
• Week four, sketch a design first, then build the sketch.
• Throughout, photograph results so progress is visible.

Perfectionism is often underneath

Frequently the refusal to improvise is fear of producing something visibly imperfect. If a child abandons any model that goes wrong, the issue is not creativity but tolerance for a bad first attempt. Building deliberately rough prototypes alongside them, and cheerfully knocking your own down, does more than encouragement does.

What to buy, and what not to

The instinct is to buy a bigger box so there is more to follow. That mostly buys a longer booklet and a bigger tidying job. If anything, buy a component type the household lacks, such as gears or wheels, because a new capability suggests new questions. And be honest with yourself about the phrase on the packaging: no set develops engineering thinking on its own, whatever the label promises. What develops it is somebody building alongside the child and asking why.

Leave it alone if it is working

A child who happily builds supplied models for two years is not failing at anything. Plenty of adults follow patterns in knitting, recipes in cooking and plans in woodwork for a lifetime and enjoy it thoroughly. Offer the bridge to free building, make it easy to cross, and drop it if the answer is a settled no.