How Kids Learn Science
Children are natural scientists. Before they set foot in a classroom, they’ve already developed elaborate theories about how the physical world works — why objects fall, how plants grow, what makes something alive. Those theories are often wrong in ways that matter. But they’re never random. They come from experience, observation, and reasoning.
Understanding how children actually form and revise scientific knowledge — rather than how we wish they would — is the foundation of effective science education. The research tells a clear story. It’s just often ignored in practice.
Children come with theories already formed
Decades of research in cognitive development and science education converge on one finding: children don’t arrive as blank slates. They arrive with prior knowledge frameworks — informal theories about how the world works — that are robust, coherent, and resistant to change.
These frameworks aren’t quirks to be corrected. They are the architecture that new understanding has to be built onto. Information that doesn’t connect to an existing framework tends not to stick. Information that directly contradicts an existing framework tends to be rejected or reinterpreted.
This is why telling a child something different from what they already believe rarely changes what they believe. Conceptual change requires more than information. It requires an experience that makes the existing framework feel insufficient — and a new framework that does a better job of explaining what just happened.
Why experience comes before explanation
The most durable learning in science happens when students encounter a phenomenon before they’re given the explanation for it. When explanation comes first, students listen and nod and remember the words — but their underlying framework doesn’t change. When experience comes first, students bring their existing framework to the situation, notice where it fails, and arrive at the discussion genuinely needing a better one.
This is why the sequence matters so much. Explore, then discuss. Discover, then explain. The game or experiment isn’t an illustration of a concept the teacher has already introduced. It’s the event that makes the concept worth having.
Professor Kasumovic’s research at UNSW consistently demonstrates this effect. In sessions with university students using Arludo, students who played before any instruction naturally produced well-formed hypotheses and were comfortable discussing their reasoning afterwards — without being prompted. The experience created the need for the language.
The role of identity in science learning
How a child learns science is not just a cognitive question. It’s also an identity question.
Research in science education consistently shows that children’s beliefs about who does science — and whether that includes them — significantly predict their engagement, persistence, and long-term participation in science. A child who doesn’t see themselves as a science person doesn’t just disengage. They actively avoid the cognitive effort that science learning requires, because the effort doesn’t feel worth it.
This effect is measurable from primary school. Children as young as eight have already formed views about whether science is for them. Those views are shaped by what scientists look like in the media, who gets celebrated in the classroom, and whether the child has experienced the feeling of genuinely figuring something out for themselves.
That last factor is the most important and the most actionable. Self-efficacy — the belief that you are capable of doing something — develops through mastery experiences. A child who has actually figured something out, who has tested a hypothesis and had it confirmed by evidence, has proof that they can do science. That proof is much more durable than being told that science is for everyone.
Social context and science learning
Children don’t learn science in isolation. They learn it in classrooms, at home, and in the context of relationships with teachers and parents who communicate — through their questions, their reactions, and their own engagement — how much scientific thinking is worth doing.
A teacher who responds to a wrong hypothesis with curiosity rather than correction is building something different from a teacher who marks it wrong and moves on. A parent who asks ‘what did you figure out?’ rather than ‘how did you go?’ is positioning the child’s reasoning as the thing worth discussing.
These interactions accumulate. A child who experiences scientific thinking as genuinely valued — whose hypotheses are taken seriously, whose questions are treated as interesting — develops a different relationship to science than a child who experiences it as a performance of right answers.
What this means for parents and teachers
The research points to three things that consistently make a difference.
First: give children experiences before explanations. Let them encounter phenomena and form their own theories before offering the authoritative version. The discussion will be richer and the understanding deeper.
Second: treat misconceptions as interesting rather than wrong. A child who says the heavier object should fall faster is doing science — they’re reasoning from evidence, even if the evidence is incomplete. The goal is not to correct the reasoning. It’s to extend it.
Third: make sure children experience the feeling of figuring something out. Not just completing tasks, but genuinely discovering something. That experience — repeated — is what builds the identity of a scientific thinker. Everything else follows from it.
About the Author
Professor Michael Kasumovic is an evolutionary biologist at UNSW Sydney and the founder of Arludo. His research explores how social interactions and playing video games alter how people perceive themselves — and how that shapes their behaviour. He has used Arludo in his own university teaching for 10 years and built the platform to turn that research into something kids, teachers, and parents actually want to use together.