The Analogy Trap: How Science Teaching Shortcuts Mislead Students
Flawed analogies create persistent mental models that coexist with correct knowledge, causing students to default to the wrong framework when solving novel problems.
Research on conceptual change reveals that students who learn through flawed analogies retain the analogy's logic alongside correct definitions, defaulting to the analogy when faced with novel problems. Every analogy contains 'non-alignable differences'—features that smuggle in misleading implications not corrected by later instruction. The solution is not to ban analogies but to treat them as temporary scaffolds with explicit removal plans: state the analogy, map its boundaries, introduce the target concept in its own terms, then deliberately dismantle the analogy.
The education orthodoxy holds that when students struggle with an abstract concept, any analogy is better than none. A messy room for entropy. A highway for electric current. A filing cabinet for computer memory. The instinct is generous: give the learner something concrete to hold onto. But the research on conceptual change tells a less comfortable story. John Clement's 1993 study on bridging analogies documented something that should unsettle every science teacher who has reached for a quick comparison. Students who learned physics through a flawed initial analogy did not simply discard it when the correct model was introduced. They retained the analogy's logic alongside the correct definition, and when faced with a novel problem, defaulted to the analogy. The wrong mental model anchored itself, and the correct one floated on top like a veneer that cracked under pressure. The mechanism is straightforward. An analogy works by aligning relational structure between a familiar domain and an unfamiliar one. The problem is that alignment is never perfect. Every analogy has what Dedre Gentner and Linsey Smith call "non-alignable differences" — features of the source domain that have no counterpart in the target. The messy room analogy for entropy maps disorder onto disorder, which is fine at a surface level. But it also smuggles in the implication that entropy is about physical clutter, that it can be reversed by tidying up, that it is a property of arrangement rather than a statistical measure of microstate distributions. Those smuggled features are not corrected by later instruction; they coexist with it. The implication is not that analogies should be banned. That would be throwing out a genuinely useful scaffold. But the scaffold must be treated as temporary and, crucially, must come with an explicit removal plan. Here is a concrete protocol drawn from Clement's work and the broader conceptual change literature: First, state the analogy and let students work with it. Second — and this is the step most curricula skip — explicitly map the boundaries of the analogy: what aligns, what does not, and where the analogy will break if pushed too far. Third, introduce the target concept in its own terms, using the analogy only as a point of departure. Fourth, return to the analogy after the concept is established and dismantle it: "Here is where the messy room analogy served us, and here is where it would mislead us." A teacher who walks through that sequence is not using an analogy as a shortcut. They are using it as a temporary bridge, and then teaching students how to leave the bridge behind. That is harder than reaching for a vivid comparison and moving on. But it is the difference between a student who can parrot a definition and one who can actually reason with the concept.