Holding Ideas in Your Hand: The Case for Concrete Manipulatives in Science Classrooms

A guide for science teachers and curriculum leaders who want students to do science, not just learn about it

I was visiting a chemistry class and was lucky enough to introduce the idea of valence electrons. The simplicity of how chemistry is organized around patterns in outer electrons has always impressed me as a science teacher. Giving students insight into how the world works is so fun. The class was talented but most didn’t have a conceptual understanding of how chemical reactions  work. I’m sure they could balance and identify complex reactions but they didn’t know what was happening at the electron level. 

They understood the words. They didn’t understand the thing.

I did a mini-lesson for the class on a simpler reaction; the formation of table salt from sodium metal and chlorine gas. I used magnetic whiteboard blocks to build the atoms and show how they combined. You could see the understanding in their eyes. When they started modeling the reactions themselves, learning spread across the room. 

That was it. Not a new explanation. Not a better diagram. The class needed to build the reactions with their hand before they could hold the ideas in their head. 

I've taught science for a long time, and I've watched that same moment happen dozens of times with dozens of students, across chemistry and biology and physics. The content varies. The mechanism is always the same. Understanding doesn't get installed from the outside. Students build it. And they build it faster, and more solidly, when they have something physical to work with.

What Modern Science Education Actually Asks Students to Do

Science education has changed. Whether your state uses NGSS, Texas TEKS, or its own framework, the direction is consistent: students shouldn't just accumulate science content. They should practice science. Ask questions. Build and revise models. Construct explanations from evidence. Argue about data. Figure out what's actually going on in a phenomenon rather than being told.

This is harder to teach than it sounds, and the hardest parts have nothing to do with content knowledge.

Take modeling. It's a core practice in every serious science framework, and for good reason. Scientists build models constantly; climate models, molecular models, population models. The model is how scientists think. But there's a version of "modeling" that happens in classrooms that isn't really modeling at all: students copy a diagram from the board into their notes. That's transcription. The student who copies the carbon cycle diagram has not built a model. They've reproduced one.

Real modeling means making decisions. What goes in the model and what gets left out? Where's the boundary of the system? Which connections matter? A student who has to answer those questions, who has to place a piece somewhere and defend where they put it, is doing something cognitively different from a student who is filling in blanks on a worksheet.

The other piece that's genuinely difficult to teach is systems thinking. All of science is ultimately about systems: atoms in molecules, organisms in ecosystems, forces in mechanical systems, variables in experimental systems. But students don't naturally see systems. They see lists. They see the parts. The connections between parts, the hierarchy, the idea that removing one piece changes the behavior of the whole requires a kind of spatial, relational thinking that most students have not been asked to do before. You can lecture about it until you're out of breath. Students develop it by building systems, not by hearing about them.

The Manipulative Gap

Third grade teachers know something that high school science teachers have largely forgotten: students learn abstract ideas through concrete experience first. You don't teach fractions by defining numerators and denominators. You cut an apple. You fold a piece of paper. You use fraction tiles until the abstraction has something to land on.

Elementary classrooms are full of manipulatives. Then students hit middle school and most of them disappear. By high school they're almost entirely gone, replaced by textbooks and slide decks and, more recently, individual devices. The assumption seems to be that older students can handle abstraction directly. Some can. Most can't, at least not without a concrete foundation underneath.

The irony is that the science gets more abstract as the manipulatives get scarcer. A third grader learning about states of matter is dealing with things they can see and touch. A tenth grader learning about electron configuration is dealing with something invisible, counterintuitive, and genuinely hard to picture. That student needs something physical to hold onto more than the third grader does, not less.

Screen-based tools fill some of this gap but not all of it. I'm not making an argument against technology in science class. I've spent years making educational videos and I believe a good simulation can do things a physical model can't. But when a student interacts with a digital simulation, they're exploring someone else's model. The parameters are already set. The variables are already chosen. The student is a visitor. When a student builds a physical model, they're the architect. That's a different cognitive experience, and it produces different learning.

There's also a collaboration problem that screens create and physical tools solve. When thirty students are each on a Chromebook, they are thirty isolated individuals moving through thirty separate experiences. There is no shared object to argue about, no common surface to point at, no moment where one student moves a piece and everyone else reacts. Physical models built on shared whiteboards create exactly that situation. The model is visible to everyone. It invites challenge. It makes disagreement productive. A lot of the best scientific thinking I've watched in classrooms happened when one student moved a block and another student said "wait, that's not right."

What Systems Thinking Looks Like When It's Working

A few years ago I watched a group of biology students model an energy flow in an ecosystem for the first time. We were in Wisconsin so we used the phenomenon of maple syrup. They used Switch-Its and a big whiteboard. The task was to show how energy and matter move through the system. The students had deep knowledge of maple syrup (e.g. seasons, grade, techniques, etc.) but they hadn’t thought it through at the level of matter and energy. One of the first steps was determining the system and the important components. They were able to collaborate with their group members and then grab knowledge from other groups. The nice thing about the blocks is that they could add, remove or modify components without drawing the whole model over again. One student moved a piece connecting sunlight to the tree and another immediately said, 'wait, but what about the freeze-thaw cycle?' That exchange wasn't in the lesson plan. The model generated it.

Learning took place socially. None of that was in the lesson plan. The model generated it. When students can see a system laid out in front of them, questions emerge that wouldn't have occurred to them from reading a description. "Where does the matter come from? What makes the sap run? How will climate change affect syrup production?" aren’t questions most students ask from a textbook passage. It's a question that jumps out when you're looking at a physical map of the connections and you can see exactly how many things depend on that one piece.

That's systems thinking. Not the vocabulary of it. The actual habit of mind. Holding the whole structure in view, reasoning about dependencies, asking what changes when something changes. Students don't develop it by having it explained. They develop it by doing it, repeatedly, with something they can see and touch and argue about.

The same dynamic shows up in chemistry when students model how electrons move between molecules. In physics when they build force diagrams that they can actually rearrange. In biology when they map the relationships between structures and functions rather than just labeling them. Physical modeling is not a special technique for one kind of content. It's a general tool for making abstract systems concrete enough to think about.

Why the First Model Should Always Be Wrong

One thing I've noticed over years of watching students build models: they're uncomfortable being wrong in front of each other. They want to produce the right answer. They've been trained to produce the right answer. A blank piece of paper in front of a class that's watching feels like a very exposed place to make a mistake.

Physical models change that dynamic. The blocks are explicitly temporary. Nothing is committed. The whole point is to put something up and then figure out what's wrong with it. When students understand that the first model is a draft - not a final answer, not something to be graded, just a starting point for an argument - the conversation opens up. Students who would never raise their hand to give a wrong answer will confidently move a block to the wrong place and defend it, because they can see everyone else's wrong models right next to theirs.

This is actually how science works. Models are always wrong in some sense. Every model is a simplification that leaves things out, emphasizes some features at the expense of others, and breaks down at the edges. Scientists don't try to build the right model on the first try. They build a model, test it against reality, find where it fails, and revise. When students do that with physical models they're not just learning science content. They're learning what science is.

For the Curriculum Directors Reading This

I want to say something directly to the curriculum leaders and instructional coaches who work with science teachers, because this decision often lives at your level.

Physical modeling materials are not elementary school supplies that students outgrow. The research on concrete-to-abstract learning applies at every grade level. If anything, the case for physical modeling gets stronger as the content gets more abstract, and high school chemistry and physics and biology are very abstract. Students who never get to touch what they're learning about are being asked to hold complex invisible systems in mind with nothing to anchor the ideas.

The other thing worth naming: when teachers are expected to implement inquiry-based instruction but only have websites, textbooks, worksheets, and individual devices to work with, inquiry becomes a performance. Teachers can ask the right questions and use the right language and still have students who are fundamentally passive, waiting to be told what the right answer is. Physical modeling materials make it structurally hard for students to be passive. Someone has to put a piece somewhere. That act starts the thinking.

Good science instruction is not about having the right curriculum. It's about giving students experiences that produce understanding. Physical tools are part of that. Schools that take them seriously, all the way through high school, get different results.

The Bottom Line

The chemistry students didn't need another explanation of valence electrons. They had plenty of explanations. They needed to break a bond with their hands, watch the pieces separate, and rebuild them into something new. The moment they did that, the explanation finally had somewhere to land.

There's a lesson in that for how we think about science instruction generally. Understanding is not a passive transfer from teacher to student. It's something students build, and they build it through experience. The question worth asking in any science classroom is whether students are getting the experiences that actually produce understanding, or whether they're getting good explanations that never quite land.

Physical models are not the whole answer. They're a starting point. But they're a starting point that a lot of science classrooms are missing, and the gap is visible.

 

AI Transparency Disclosure: This content was created with the assistance of artificial intelligence. While AI helped with drafting based on provided topics, the final version has been reviewed, edited, and approved by a human author who takes full responsibility for its accuracy and perspective.