What a Physics Answer Can Tell Us
Two students write the same answer for the period of a pendulum. One begins by specifying a small angle, a nearly rigid length, and negligible friction. The other recalls a formula and inserts the numbers. Their arithmetic may be identical. The evidence of their physical reasoning is not.
An equation is among physics’ most powerful tools. The educational problem arises when a correct numerical result becomes the whole record of learning. If we want a learner to use physics in an unfamiliar situation, the task has to reveal how the learner selected the model, interpreted its terms, and decided whether the result applies.[1,2]
In What a Simulation Can Tell Us, I asked which part of a simulated prediction has been checked: the code, the numerical solution, the physical model, or the proposed explanation. The parallel question for a classroom is direct: which part of that reasoning can the student defend? A plotted curve or a final number cannot answer on the student’s behalf.
A formula has a domain
Consider a simple pendulum. Under an idealized small-angle approximation, its period is T₀ = 2π√(L/g), where L is the pendulum length and g is the gravitational acceleration. The formula predicts that the period increases with √L. For this ideal system at small amplitude, the period does not depend on the bob’s mass. These claims follow from a model with stated assumptions, not from the look of a familiar symbol.
To make the modeling visible, ask learners to identify the system and draw the angle from the vertical. Before any calculation, ask what would change if the length doubled, or if the release angle became much larger. The first question tests how the equation represents length. The second tests whether the learner notices the small-angle condition. At larger amplitudes, the ideal pendulum takes longer than the small-angle approximation predicts; a real classroom apparatus may also show friction and measurement effects.
The teacher can model the initial steps. Learners need enough knowledge of forces, motion, radians, and algebra to reason with the expression. The point is to connect those prerequisites with model choice and interpretation. The National Research Council’s science framework likewise treats knowledge and scientific practices as intertwined rather than as rival aims.[1]
A solution has several parts
A written solution may contain a useful representation, a suitable physical principle, a correct application to the particular conditions, valid mathematics, and a logical path between them. Docktor and colleagues developed and studied a rubric for these different components in introductory university physics.[2] A mark awarded only for the last line misses distinctions that the rubric can expose.
That does not mean every assignment needs a long essay. A brief prompt might ask: “Which approximation did you use? Predict the direction of the error at a larger release angle. What observation would test that prediction?” A concise answer can reveal more about the learner’s model than another substitution exercise. An incorrect answer can also be diagnostic: a learner who chose the right model but slipped in arithmetic needs different feedback from one who applied the wrong model flawlessly.
Physics education research has also developed formative tasks and rubrics for experimental abilities.[3] In a transformed electronics laboratory, a study of student notebooks found that making the comparison between interpreted data and model prediction was particularly difficult. More scaffolded activities in that setting supported fuller engagement with the modeling process.[4] This is a reason to teach the comparison explicitly, not proof that one identical scaffold works for every age or topic.
Ask what the evidence tests
Return to the pendulum. A student may run a simulation, obtain a smooth trajectory, and measure oscillations. I would ask the student to label three different activities: checking the numerical calculation, fitting a parameter using selected observations, and comparing a prediction with a separate observation. Their Week 9 meanings should remain distinct. A small-angle simulation that agrees with small swings is valuable, but the test does not silently extend to large swings.
For a classroom task, provide a small set of measurements with their units and uncertainty, and identify which were used to set model parameters. Then ask what additional measurement would challenge the model. If a data set is synthetic, label it synthetic. If actual measurements are unavailable, assess the proposed test as a plan rather than reporting validation. Students can learn to say “the evidence supports this model for these conditions” without turning an incomplete comparison into absolute proof.
The five moves I recommend are: represent the system; state a prediction; calculate with a justified model; compare against relevant evidence; revise or bound the claim. This is my instructional synthesis of the cited literature and the Week 9 analysis. Its effectiveness as a particular sequence has not been established by the sources cited here. It offers a visible structure for a teacher to adapt and evaluate.[1–4]
Assess transfer, with care
One successful explanation on a familiar pendulum does not establish independent reasoning. After feedback, change the surface and the physical demand: ask about a mass–spring oscillator, a pendulum with a moving support, or a case where friction cannot be ignored. Specify the knowledge provided so that the new question tests a defined ability rather than simply rewarding students who already know a new formula. A delayed task can check retention separately from immediate performance.
There is a caution in the evidence. Kuo and colleagues found that a targeted approach increased certain uses of mathematical sensemaking and correctness on tailored crossover items, but it did not increase correctness on related standard problems.[5] A 2026 study across five German secondary schools showed how a rubric and conceptual test could reveal different problem-solving patterns; its 51 participants and observational design do not establish broad causal effects.[6] The outcome measure matters. So do age, preparation, task difficulty, language, and scoring reliability.
This is why I would keep two things visible in assessment: the numerical result and the reason it deserves trust. Formula fluency matters. A learner also needs to know when the formula has stopped answering the question. For educators, the next useful step is modest: take one existing problem and add a model assumption, an evidence question, and a changed-condition check. Compare the work students actually produce before claiming improvement.
Short author bio. Dr. Albert Tan Lie Sing is a mathematical physicist, AI robotics systems architect, and STEAM education innovator. His work connects physical modeling, intelligent systems, and evidence-led learning.
Connecting Articles: What a Simulation Can Tell Us; and Prediction Has Limits.
References
1. National Research Council. (2012). A framework for K–12 science education: Practices, crosscutting concepts, and core ideas. National Academies Press. https://doi.org/10.17226/13165
2. Docktor, J. L., et al. (2016). Assessing student written problem solutions: A problem-solving rubric with application to introductory physics. Physical Review Physics Education Research, 12, 010130. https://doi.org/10.1103/PhysRevPhysEducRes.12.010130
3. Etkina, E., et al. (2006). Scientific abilities and their assessment. Physical Review Special Topics—Physics Education Research, 2, 020103. https://doi.org/10.1103/PhysRevSTPER.2.020103
4. Stanley, J. T., Su, W., & Lewandowski, H. J. (2017). Using lab notebooks to examine students’ engagement in modeling in an upper-division electronics lab course. Physical Review Physics Education Research, 13, 020127. https://doi.org/10.1103/PhysRevPhysEducRes.13.020127
5. Kuo, E., Hull, M. M., Elby, A., & Gupta, A. (2020). Assessing mathematical sensemaking in physics through calculation-concept crossover. Physical Review Physics Education Research, 16, 020109. https://doi.org/10.1103/PhysRevPhysEducRes.16.020109
6. Meyer, A., Fischer, S., & Friege, G. (2026). Analyzing problem solving in secondary physics education: A rubric-guided approach to explore individual learning needs. Physical Review Physics Education Research, 22, 010111. https://doi.org/10.1103/3bs7-fnrd
Comments