Timed classification challenge
The student is handed two or three unfamiliar specimens and a clock. Working against time, they examine each animal’s observable traits — symmetry, segmentation, appendages, skeleton — and walk a dichotomous key couplet by couplet, deciding which trait to check next based on what the last one showed. At the end they place each specimen in its group and justify it from the features on the bench. There is nothing to copy and no key to the answer: the specimens are real, the time is real, and the classification has to hold up.
| Criterion | Not yet | Approaching | Mastered |
|---|---|---|---|
| Systematic procedure | Examines traits at random and loses track of which specimen is which. | Has a rough plan but skips a confirming trait or re-checks traits inefficiently. | Follows a logical sequence, letting each trait decide the next couplet, and tracks every specimen without confusion. |
| Reading observable traits | Cannot connect a trait to a group, or overlooks the diagnostic feature. | Reads obvious traits (legs, shell) but hesitates on close calls like arachnid vs. insect. | Reads diagnostic traits confidently — symmetry, segment count, leg number, presence of a backbone — and knows which ones actually separate the groups. |
| Working the dichotomous key | Guesses instead of following the key, or takes the wrong branch and never checks. | Follows the key but misreads a couplet or forgets to confirm the endpoint. | Works the key couplet by couplet, choosing each branch from an observed trait, and confirms the result before committing. |
| Accuracy under time pressure | Rushes into misreads, 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 classification. |
| Classification & justification | Guesses a group unsupported by the traits observed. | Names the group but cites only one trait, or hedges between two. | Places each specimen in its group and justifies it by chaining several observed traits into a single conclusion. |
“Eight legs, two body regions, no antennae — that’s an arachnid, not an insect, so it keys to a spider, not a beetle. I checked the leg count and the body plan before I committed, because the antennae are the trait that actually splits the two.”
“It’s got a bunch of legs, so it’s a bug of some kind. I’d have to guess which group without looking it up.”
This assessment is AI-proof by design: it happens at the bench, with real specimens, against a real clock. No chatbot can turn a specimen over, read a trait it was not handed, or decide the next couplet while the timer runs. The specimens differ from student to student, so there is no answer to look up — mastery is shown by keying out and classifying in person, not by submitting.