Timed oceanographic data reading
The student is handed a packet of real ocean data and a clock. Working against time, they read temperature–salinity–depth profiles, tide tables, plankton counts, dissolved-oxygen profiles, and a species-distribution map — pulling the right values, with units, and naming the pattern behind each one. Then they draw a conclusion the numbers actually support. There is nothing to copy and no key to consult: the data is real, the time is real, and the reading has to hold up.
| Criterion | Not yet | Approaching | Mastered |
|---|---|---|---|
| Reading graphs & tables | Misreads axes or columns and cannot tell which line or column is which. | Reads a simple graph but confuses axes on a two-variable plot or misaligns a tide-table row. | Reads temperature–salinity–depth profiles, tide tables, and distribution maps accurately, tracking the right line, column, and axis every time. |
| Extracting values with units | Pulls numbers off the page with no units, or reads the wrong point entirely. | Extracts values but drops units or interpolates carelessly between gridlines. | Extracts the right value at the right depth or time and reports it with correct units — °C, salinity, metres, mg/L. |
| Interpreting patterns | Sees a wiggly line but cannot name the pattern behind it. | Spots an obvious feature but mislabels it — calls a tidal cycle a trend, or misses the thermocline. | Names the pattern from the data — the thermocline in a depth profile, the tidal cycle in a tide table, a plankton bloom, an oxygen minimum zone. |
| Drawing a defensible conclusion | States a conclusion the data does not support, or none at all. | Reaches a conclusion but cannot point to the values that back it. | Draws a conclusion the data supports — and cites the specific readings that lead to it. |
| Accuracy under time pressure | Rushes and makes reading errors, or freezes and runs out of time. | Finishes most of the task but grows careless as the clock runs down. | Works steadily against the clock — deliberate, accurate reads with time left to check the answer. |
“Temperature holds near 22°C down to about 40 metres, then drops sharply to 8°C by 100 — that steep band is the thermocline. Below it the water’s cold and dense, so the two layers barely mix. I read those depths straight off the profile before I called it.”
“The line goes down, so it’s getting… colder deeper? I’m not sure what to call that part. I’d need more time to figure out the units.”
This assessment is AI-proof by design: it happens in the room, with a real dataset, against a real clock. No chatbot can find the thermocline in a profile it was not handed, read a tide table under time pressure, or defend a value it did not extract. The packets differ from student to student, so there is no answer to look up — mastery is shown by reading and interpreting in person, not by submitting.