Hands-On Skills: What Wooden Puzzles Teach Beyond Solving

A meta-analysis of 20 spatial skills training studies with young children reported an average effect size of 0.96, which is large by the standards of educational research. That is a real finding and it is worth taking seriously. It is also narrower than the claims usually made for puzzles, and the gap between the two is what this post is about.

What the training evidence actually shows

The meta-analysis by Yang and colleagues, Is Early Spatial Skills Training Effective?, published in Frontiers in Psychology in 2020, pooled 20 intervention studies from 2009 to 2020 covering children aged 0 to 8, producing 50 independent effect sizes from 1,535 participants. It concluded that diverse training strategies including hands-on exploration, visual prompts and gestural spatial training significantly foster young children's spatial skills, with an overall Hedges's g of 0.96.

Two things need saying immediately. First, spatial skills are trainable, which was not always assumed. Second, the authors are explicit that they did not include studies on transfer of spatial skills training to untrained tasks, and they list several limitations: a relatively small number of studies with high heterogeneity, exclusion of non-English publications, insufficient reporting of trainer qualifications and intervention duration, and no longitudinal data on long-term effects.

So the honest summary is that practising spatial tasks improves performance on spatial tasks, in young children, in the short term, and the question of what carries over to anything else was not what this analysis measured.

What hands find out that eyes cannot

The more distinctive thing about a wooden puzzle, as opposed to a screen version, is that you handle it. Touch is not a degraded version of vision. It gathers different information, and it does so through specific, stereotyped movements.

Research by Klatzky, Lederman and Reed published in the Journal of Experimental Psychology in 1989 on haptic integration of object properties examined how people classify objects by touch across texture, hardness and planar contour. They found a redundancy gain when a second property was available, and that texture and hardness were integrated during exploration so strongly that removing either one slowed responses, even when participants had been told to attend to only one. Combining texture or hardness with contour showed less integration, and the authors attribute this to the compatibility of the exploratory movements involved.

The practical translation for a puzzle solver: the hand movement you use to judge whether a face is smooth is not the same movement you use to trace an edge, and you cannot do both at once. Experienced solvers of blind or seized puzzles do these things in sequence, deliberately.

Sequencing and rejecting bad options

Interlocking puzzles reward two habits that are not about visualisation at all: doing things in order, and abandoning wrong ideas quickly.

The second turns out to be measurable. A 2025 study in the Journal of Intelligence on object assembly tasks and spatial ability by Li, Man and Lakin broke performance into five attributes, including encoding, mental rotation, mental displacement, and falsification, meaning the rejection of options with obvious mismatches. Working with 50 children aged 6 to 14, they found mastery of falsification was the strongest predictor of success, ahead of mental rotation.

That fits what anyone who has spent an evening with a burr will recognise. The person who finishes is not usually the one with the best mental picture. It is the one who stops pursuing a dead line after three moves rather than thirty.

Being honest about the limits

• Improvement on spatial tasks does not automatically mean improvement in mathematics, engineering, reading, or general intelligence. Transfer has to be demonstrated separately, and the meta-analysis above deliberately excluded it.
• Effects measured in young children do not necessarily apply to adults.
• Short-term gains in a study are not the same as durable skills. Longitudinal evidence is thin.
• Puzzles are not a treatment. They do not prevent, treat, or cure any condition, and anyone with concerns about cognition, development or dexterity should speak to a doctor or an occupational therapist rather than buying a puzzle.

How to frame this as a parent or teacher

You do not need the inflated version to justify wooden puzzles. The defensible framing is simpler: they give children repeated, self-directed practice at manipulating objects in space, they provide immediate physical feedback without a score or a timer, and the evidence that this kind of practice improves spatial performance is reasonably good.

Beyond that, they are absorbing, they last decades, and they are one of the few activities where a child and an adult can be genuinely equally stuck. That is worth something on its own, with no research citation required.