Timed prediction-and-test
The student is handed a setup and a clock. Working against time, they take a measurement and use it to predict what will happen — how far a cart will roll, how long a pendulum will swing, whether a beam will balance — and write the prediction down before they run it. Then they run it live, compare what happened to what they said, and account for the gap. There is nothing to copy and no key to consult: the setup is real, the time is real, and the prediction has to be committed to out loud before the outcome is known.
| Criterion | Not yet | Approaching | Mastered |
|---|---|---|---|
| Reasoned prediction | Guesses an outcome with no measurement or reason behind it. | Makes a prediction but cannot tie it to the measurement or a rule. | Predicts the outcome from the measurement and explains the reasoning before running it. |
| Procedure under time | Runs the test at random or forgets steps as the clock runs. | Has a rough plan but skips a step or repeats work when rushed. | Follows a clear, ordered procedure and stays on track even under the clock. |
| Measurement & technique | Reads the tool wrong or records numbers with no units. | Generally careful but gets sloppy as time runs — misreads a scale or crowds the setup. | Reads every tool with the right units and keeps the setup clean and controlled throughout. |
| Honest accounting of the gap | Ignores the difference between prediction and result, or hides it. | Notices the gap but cannot say what caused it. | Names the gap between prediction and result honestly and reasons about the cause — friction, a timing error, a rough measurement. |
| Oral defense under questioning | Folds at the first follow-up or recites a line that does not fit the run. | Answers some follow-ups, falters when asked to justify a choice or a number. | Handles unrehearsed follow-ups about this run with sound, on-the-spot reasoning. |
“I measured the ramp height and used it to predict the cart would reach the line in about two seconds. It actually took two-and-a-quarter — slower than I said, because the ramp has more friction than a perfect one. I wrote my prediction down before I let go, so I couldn’t fudge it after.”
“I said it would go fast, and it kind of did. I don’t really know why it was different from what I guessed — I didn’t write the guess down first.”
This assessment is AI-proof by design: it happens at the bench, with a real setup, against a real clock. No chatbot can commit to a prediction, run it live, and account for the gap while the timer runs. The setup differs from student to student, so there is no answer to look up — mastery is shown by predicting and defending in person, not by submitting.