Physics problems are not mainly tests of formula memory. They are tests of model selection. Before an equation is useful, you need to decide what system you are describing, which interactions matter, what can be neglected under the question’s assumptions, and how a measurable quantity follows from that model. A student who starts by hunting through a formula sheet may still reach the right answer on a familiar question. A student who starts with a model is far more likely to handle a changed context. This guide is a practical routine for mechanics and introductory physics, but the habits transfer to electricity, waves, thermal physics, and beyond. The focus is on making reasoning visible: diagrams, statements of assumptions, units, estimates, and short explanations. It does not promise a particular mark; it gives you a way to identify why an answer failed and what to practise next. That distinction matters because correct terminology can mask an incorrect model. The Force Concept Inventory was designed around common conceptual difficulties in introductory mechanics. Its lasting lesson for learners is modest but important: being able to state a law is not the same as being able to use that law to predict motion in a concrete situation. Begin every problem with a model statement Write four lines before touching a formula: System: What object, collection of objects, field region, or volume is being analysed? Known and sought: Which quantities are given, which are unknown, and which direction/sign convention will be used? Interactions: What forces, energy transfers, fields, constraints, or waves act on the system? Assumptions: Is drag negligible? Is a string light and taut? Is the field uniform? Is energy loss ignored? Is motion one-dimensional? These lines take less than a minute and prevent many wrong starts. For a block pulled across a rough surface, “the system is the block” immediately clarifies that the pull, normal contact force, gravity, and friction are external forces on that system. For an energy calculation involving a block and an Earth–block system, you may define a different system because gravitational potential energy is now internal to the chosen system. Neither choice is automatically correct; the point is to make the choice and use it consistently. A model statement also makes your work checkable. If you later discover that friction cannot be ignored, you know which assumption has to be revised. Without a written model, students often patch formulas together and cannot locate the decision that created the error. Draw diagrams that carry information A physics diagram is a compressed argument, not a decorative sketch. In mechanics, a free-body diagram should contain only the forces acting on the chosen object, each drawn from or associated with that object. Do not include a force the object exerts on something else on the same diagram. Label direction and, when useful, choose axes aligned with a ramp, field, or motion constraint. Separate different diagrams when they answer different questions. A motion diagram tracks position or velocity over time. A free-body diagram tracks interactions. An energy bar chart tracks system energy changes. A circuit diagram tracks connections, not the physical placement of wires. Combining all of them into one busy drawing often hides the very distinction you need to reason clearly. After drawing, narrate it in one sentence: “The only horizontal force opposing the motion is kinetic friction, so the horizontal acceleration must be opposite the velocity if no larger forward force is present.” This sentence is an early prediction. If your later algebra says acceleration has the opposite sign, you have a conflict to resolve rather than an answer to submit. Choose the principle before choosing the formula Most introductory problems are built around a small set of principles: Newton’s laws, conservation of energy, momentum conservation, impulse, kinematic relations under stated acceleration conditions, circuit laws, or wave relations. The same symbols can occur in several of them, so a variable list is not a method. Ask three diagnostic questions: What is changing? Velocity, position, energy form, momentum, charge distribution, pressure, or wave phase? What interaction accounts for that change? A net force, external impulse, work, thermal transfer, electric field, or a boundary condition? What is conserved within the defined system? Be precise: momentum may be approximately conserved while mechanical energy is not, or charge may be conserved while current differs in separate branches. For example, a collision question does not automatically call for energy conservation. If external impulse is negligible over the collision interval, momentum conservation may be the useful starting point. If the objects stick together, kinetic energy is not conserved, although total energy remains accounted for in a wider description. Saying “energy is always conserved” is true but too broad to solve the problem until you identify the system and the energy pathways. Work through one situation in layers Imagine a cart released from rest at the top of a track, then moving along a rough horizontal section. First sketch the path and choose a height reference. Next decide whether the system is the cart alone or cart plus Earth plus track. If the system includes the track and friction is present, some mechanical energy may be transferred to internal energy; if you choose the cart alone, contact forces can do external work. Both descriptions can be valid, but they organize the bookkeeping differently. Now make a qualitative prediction before calculating: the cart’s gravitational potential energy decreases on the descent; its kinetic energy increases unless another transfer is significant; the rough section reduces its kinetic energy. Only after that story is coherent should you write an energy equation with the terms that belong to your stated system. Finally check the result: a negative speed, a final height above the release height without an external input, or a unit of kilogram-metres per second where an energy was requested means the model or algebra needs review. This layered approach feels slower at first. It becomes faster because it replaces blind substitution with reusable decisions. In research on physics problem solving, Heller, Keith, and Anderson studied structured cooperative problem solving in an introductory course; their work is one useful historical source for treating problem solving as a learnable process rather than an innate talent. Read the 1992 study for the original context rather than assuming every group arrangement produces the same result. Use equations as claims with units and limits Before substituting numbers, write each equation in words. For instance, “net force equals mass times acceleration” claims a relation for a chosen object in an inertial reference frame. “Change in momentum equals impulse” relates a force–time effect to a momentum change. The words expose missing conditions and keep you from using an equation merely because it contains the variables in the question. Then use units at three moments: before calculation, during a multi-step derivation, and after calculation. Units do not prove the physical principle, but they catch many slips. If an acceleration emerges in newtons or a time is added to a distance, stop. For symbolic manipulation, preserve units in annotations instead of deleting them mentally at the first line. Estimation is the second check. Ask whether the magnitude is plausible for the situation. A person walking at hundreds of metres per second, a household battery storing megajoules, or a mass that changes by a factor of one thousand after a centimetre-to-metre conversion are prompts to reopen the work. Estimation is not guessing; it is using physical scale to test an algebraic claim. Explain worked examples to yourself, then vary them A worked example is useful only if you actively explain why each step follows. Cover the next line and answer: “What would I do now, and what fact makes that legitimate?” If you cannot answer, uncover one line, restate it in your own words, and cover it again. This is self-explanation—not merely reading a solution more slowly. Evidence from Chi and colleagues’ study of self-explanations supports the idea that prompting explanations while studying worked examples can improve understanding. The practical boundary is important: explaining an incorrect step repeatedly does not create expertise. Compare with a reliable source and name the exact rule, model, or assumption that justifies the step. After one example, create a variation. Reverse a direction, change the system boundary, remove a stated simplification, ask for a different quantity, or present the information in a graph rather than a sentence. If your method collapses under a small change, you learned a script rather than a principle. Keep the variation modest enough that you can diagnose the new decision. Practise conceptual predictions as seriously as calculations Every formula-based question should have a companion prediction question. Examples include: What happens to acceleration if the net force doubles while mass remains fixed? Can two objects have the same velocity but different net forces? Does a constant velocity require a constant force in the direction of motion? Which circuit quantities change if one parallel branch is removed? Explain before calculating. When your prediction is wrong, do not simply replace it with the official answer. Draw the situation again and write the tempting incorrect model beside the corrected model. In mechanics, common temptations include treating motion as evidence of a forward net force, pairing action–reaction forces on the same object, or using a force diagram as though it were a velocity diagram. The value of a concept inventory is not a score; it is a vocabulary for locating these durable misconceptions. Make laboratory data part of your revision Physics is also an experimental discipline. When reviewing a practical, write what was measured directly, what was calculated, what was controlled, and what the graph could support. Distinguish random variation from a plausible systematic effect; do not claim that every imperfect point is “human error.” Consider resolution, calibration, alignment, reaction time, model assumptions, environmental factors, and data-selection choices. For a graph, state the variables and units, interpret the gradient or area only when the relationship justifies it, and explain whether an intercept has a physical meaning. A straight line can support a model within the observed range; it rarely proves a universal law by itself. This evidence-first habit helps on practical exams and makes theory less detached from measurement. Use retrieval and spacing without turning revision into a ritual At the end of a session, close the book and draw one diagram, state one principle with its conditions, and solve one short problem from memory. The objective is diagnosis. Research on test-enhanced learning found that retrieval can strengthen delayed retention relative to repeated study in the study’s setting; see Roediger and Karpicke for the original research. In physics, retrieval should include explanations and decisions, not only formula names. Return to the same idea on later days with a changed problem. The National Academies’ How People Learn II emphasizes that learning is shaped by prior knowledge, context, motivation, and opportunities to reason—not by a single universal routine. So adapt the schedule to your course and time, but keep a record of what you could do unaided at each review. A four-pass correction method Model pass: Was the system, diagram, interaction, or assumption wrong? Principle pass: Did you choose an inapplicable conservation law, equation, or condition? Representation pass: Did you misread a graph, sign, axis, vector, direction, or unit prefix? Execution pass: Was the physics sound but the algebra, arithmetic, or calculator input wrong? Record one example for each recurring category and write a repair prompt. “Draw the object and list forces before resolving components” is actionable. “I need to focus more” is not. On the next review, choose a problem that specifically tests the repair rather than repeating the identical question. The takeaway Physics becomes more transferable when you treat every answer as the end of an argument: define the system, draw the relevant representation, state the governing principle and assumptions, calculate with units, estimate the result, and explain what it means. Formulas then become concise tools inside a model rather than a catalogue to search under pressure. That is the study habit most likely to survive an unfamiliar question.