Oral lab-notebook defense
The student opens their own earth science notebook and walks the guide through one investigation start to finish: the question they set out to answer, the method they actually ran, the data as they recorded it — with units and significant figures — the anomalies they noticed, and what they concluded. The guide stops them with questions: why that mass, why those digits, what that odd reading means. Because it is the student's own handwritten record, there is no generic answer that fits — only the run they did.
| Criterion | Not yet | Approaching | Mastered |
|---|---|---|---|
| Question & purpose | Cannot state what the investigation was trying to find out. | States the question loosely but cannot tie it to a testable variable. | States a clear, specific question and explains what was being varied and measured to answer it. |
| Method & reproducibility | Cannot reconstruct what was done; steps are missing or out of order. | Recounts the main steps but leaves gaps that would stop someone else from repeating it. | Walks through a complete, ordered procedure another student could reproduce from the notebook alone. |
| Data, units & significant figures | Records bare numbers — no units, inconsistent or invented precision. | Mostly labels units but mishandles sig figs or rounds inconsistently. | Records every measurement with correct units and a precision the instrument actually supports, consistently throughout. |
| Anomalies & honesty | Hides or smooths over readings that did not fit. | Notes an anomaly but offers no account of what caused it. | Flags anomalous readings honestly and reasons about their likely cause without erasing them. |
| Interpretation under questioning | States a conclusion the data does not support, or folds at the first follow-up. | Draws a reasonable conclusion but cannot defend a specific number or choice when pressed. | Ties the conclusion back to the recorded evidence and defends each number and decision under unrehearsed questioning. |
“Here’s the entry from week three — I measured the sample’s mass and volume, noted the air bubble I trapped, and that’s why my density came out a little low. Anyone could repeat this run from these pages, including the mistake I’d fix next time.”
“I wrote the answer down at the end. The middle parts I kind of remember. I didn’t note what went wrong because it mostly worked.”
This assessment is AI-proof by design: the student defends their own handwritten notebook, out loud, in real time. No chatbot can explain a mass it did not weigh, account for an anomaly it never saw, or justify a significant figure under a follow-up question. The record is unique to the run the student did — mastery is shown by defending it in person, not by submitting it.