External empirical evidence can refute a computational model of particle interactions when the model's own simulation cannot¶

Claim · All claims
Evidence · 1 study · qualitative r3 · q1 · n=18 consented students (of 29 in the class), one sixth-grade classroom, one experienced teacher, 6-day unit
Currency · newest study 2025 · not yet checked against its sources · not reviewed by a person · recheck by 2028: tool and model accuracy changes quickly

When a sixth-grade class's computational simulation of particle diffusion could not visually distinguish between a correct theory and a plausible but incorrect "infect" theory, a physical evaporation experiment that directly tested the underlying particle-level claim shifted a substantial share of students away from the incorrect theory.

Subclaims¶

q1-q2 i? Of 18 consented students whose reasoning was tracked across a 6-day diffusion unit, 7 (39%) explicitly cited a physical evaporation experiment — in which evaporated water/food-coloring mixture turned clear, directly contradicting the prediction of an "infect" theory that ink particles chemically bond with water particles — as the specific evidence that changed their theory preference. → Wagh et al. 2025

q1-q2 i? The teacher recognized, mid-unit, that the class's own agent-based simulation could not adjudicate between the "infect" theory and the correct diffusion theory because both looked visually similar when run, and deliberately introduced the external evaporation experiment specifically because it tested a claim the simulation itself could not test. → Wagh et al. 2025

Evidence¶

Wagh et al. 2025¶

Wagh, A., Rosenbaum, L. F., Fuhrmann, T., Eloy, A., Blikstein, P., & Wilkerson, M. (2025). Toward Ontological Alignment: Coordinating Student Ideas with the Representational System of a Computational Modeling Unit for Science Learning. Cognition and Instruction, 43(1-2), 1-32. https://doi.org/10.1080/07370008.2024.2427400

q1-q2 · single-classroom qualitative observation with systematic interaction-analysis coding (three researchers, 296 coded utterances, 8/296 disagreements resolved by majority) and triangulated video/artifact data; no comparison classroom or controlled test of the evaporation experiment's specific causal effect i? · the abstract prints no effect size; the full text may n=18 consented students (of 29 in the class), one sixth-grade classroom, one experienced teacher, 6-day unit · qualitative · r3

In a sixth-grade computational-modeling unit on particle diffusion using Three Practices for Ontological Alignment in Computational Modeling, the teacher recognized that the class's agent-based simulation could not visually distinguish between two competing student theories and introduced a physical evaporation experiment whose result directly contradicted one theory's prediction. Post-unit written reflections from 18 consented students were analyzed for whether and how they referenced this experiment in explaining a change in their thinking.

Discussion¶

This is a single-classroom qualitative study without a comparison condition, so the claim documents what happened in this specific instructional sequence, not a general or causally-isolated effect of introducing external evidence. The 39% figure reflects only students who explicitly mentioned the experiment in a written reflection; it likely undercounts students whose thinking was influenced without stating so directly, and may also reflect recency bias (the experiment occurred late in the unit). No independent measure verifies that these students' post-experiment theories were durable beyond the unit's assessment.

Related Claims¶

• (none yet linked)
• In one experiment, refutational text (misbelief stated first, then fact) enhanced free recall relative to traditional text and free and cued recall relative to inverted refutational text, but its reported 14-day recognition advantage was over inverted refutational text only — related
• Students' empirical fits of simulated magnetopause locations show good internal agreement, though student estimates and simulation-calculated locations often disagree beyond uncertainties — related
• Simulation usefully explores the plausibility of models where empirical testing lags, and permits psycho-surgery experiments impossible with human subjects — related
• The claim that computer simulation necessarily enforces clearer, more rigorous psychological theory specification does not hold up — related
• Post-interviews with props and embodiment revealed understanding of macro–micro particle changes that the multiple-choice post-test did not capture. — related
• Latent transition analysis extracted four ability-profiles of hypothesis-based reasoning across affirmative, refuting, and irrelevant evidence — related
• Experts' ontological choices about the same physical phenomenon are context-dependent — related