This is a teaching record, not a grade or diagnosis. Label independent, prompted or not yet observed; record a short evidence quote and any access support. Everyone attempts the daily exit. The teacher follows 2–3 focus learners per day, rotating names; hold a later quiet conversation for a learner whose independent thinking was hidden by a pair task. Do not publish child work, recordings or names.
| Day | Keyed evidence from the lesson exit | Likely misconception | Immediate next move |
|---|---|---|---|
| 1 | “The seventh shape makes Mira check the rule because it has both smooth and jagged sides”; cited pause/retest or revised note. | “Mira is clever” with no textual detail. | Reopen one sentence; model evidence → inference, then ask for a second example. |
| 2 | A has setup/complication/ending for narrative; B defines, gives steps and states a limit for explanation. | Labels “story” and “facts” without effect. | Compare how dialogue versus imperative helps a specific reader. |
| 3 | Key is paired identification decisions, supported by “paired questions”; branch is a route after an answer. | Uses everyday lock/tree sense only. | Highlight local context clue and test a new sentence. |
| 4 | “If the card has a jagged section, follow that route. Therefore, check whether the label fits.” Condition and result are complete; connector signals consequence. | “If…” fragment or unrelated “therefore”. | Move clause strips; read sentence aloud for a complete main clause. |
| 5 | Definition, ordered route, test with all six and a limit for a new/unmatched card. An example that merely lists labels is incomplete. | Claims key always identifies any leaf or proves evolutionary relationship. | Sort model identification from real-world relationship; have a peer follow one question literally. |
| 6 | C claims A, cites $45/12 of 30 and notes poll/repair limits; D claims B, cites $50 and names access check. | Treats a proposed benefit as a measured fact. | Mark “quote/poll fact” versus “prediction” in different columns, then rewrite. |
| 7 | “In this fictional poll, 12 of 30 responses chose A; this does not establish whole-school preference.” | Generalises the small self-selected group to all students. | Ask “Who answered? Who did not?” then scope the sentence. |
| 8 | A fair alternative accurately names a benefit; colon introduces an explanation, or brackets add a short qualification. | Counterargument caricatures the other option; punctuation placed by appearance. | Read the other note aloud; replace caricature with its strongest supported reason; check colon meaning. |
| 9 | Claim, two brief facts, separate predicted benefit, own ranking and benefit of next-best forgone option, one check. | Says opportunity cost equals price or remaining cash. | Have student rank options first and articulate the benefit of rank 2. |
| 10 | See Check B below. | Overclaim, unsupported recommendation or missing limit. | Select source-scope, evidence-reason or cohesion mini-lesson as indicated. |
Check A — Day 5 explanation sample and feedback
Prompt: “Using the six model cards, explain how a two-choice key works. Include a way to test it and a limit.” No claim about real species is required.
Secure sample: “A two-choice key sends a user through paired questions about visible features. First, look at the whole edge and choose the matching description. Follow that branch until it reaches a label. Give the same cards to another person and see whether the labels match. If a new shape has no route, record it as unmatched and revise the key.” This names purpose, sequence, test and scope.
Developing sample: “Look at a leaf, pick a side and find its name. This works for every plant.” The first two imperatives are vague and the final claim exceeds what paper models can show. Feedback: “Name the feature and the set this key covers. What would you write when a new card does not fit?” Next task: rewrite two sentences and peer-test one route. If a learner gives the explanation orally, keep an exact short transcript of the evidence and note the route.
Check B — Day 10 fresh reading and writing
Original fictional stimulus: “The Friday club should run Option C. It costs $35, so it leaves $25 of the $60 supply budget. Eight of 30 poll respondents chose it. This proves all students want a mural. We still need to check display permission.” The invented poll and quote belong only to the club brief.
Student tasks: (1) identify the claim; (2) quote one supported number; (3) find an overclaim or unanswered question and repair it; (4) write a 3–4 sentence recommendation for any one option, including a fair next-best alternative, its forgone benefit and one check. Permit another justified ranking.
Key: Claim = run C. Supported numbers = $35 cost, $25 left, 8/30 responses. Overclaim = 8/30 cannot prove all students want a mural; it is fewer than one third of the responses and an unknown share of all students. “In the fictional poll, eight of thirty respondents chose C” is a fair repair. Display permission is an acknowledged check; actual access/material suitability is another possible question. If student ranks A as next best after C, the forgone benefit is potential reuse of repaired games; if B is next best, it is potential closer observation with reusable magnifiers. Neither $35 nor $25 left is itself the opportunity cost. A justified recommendation may choose A, B or C.
Feedback decisions: If number is accurate but population overgeneralised, reteach sample/scope. If claim and evidence are present but no reason links them, model “This matters because…” and ask for a specific benefit. If the next-best choice is unranked, ask for a personal decision criterion and ranking before the cost sentence. If punctuation/cohesion obscures meaning, cut and reorder the two paragraphs; do not penalise an access route that conveys the same analysis.
Recording strip
Date | Day | Focus learner code (local only) | exact evidence | independent/prompted/not yet observed | access/support | next move | revisited on. Store under school policy. Do not transmit identifiable data to SubjectNest.