# Teacher-only unseen checks, keys and next moves · mathematics Days 135 and 140

Prepare behind a cover. Record each child's first independent move/answer and actual access/prompt **before** correction. These are formative samples; do not derive a grade, diagnosis, full description attainment, science fact or real opinion from one response. Adult reads labels as needed. If a physical mass setup fails preflight, repair it before the learner sees it and record the change.

## Day 135 · fresh length and mass pairs

**Length setup:** Cut two new broad paper strips **R 45 mm** and **S 68 mm** long at actual size. Place R beginning 20 mm farther right than S, so the far ends cannot be compared by eye from different starts. Child must move or direct movement to align the *near* ends before deciding. Ask “Which is longer in this pair, and how did you check?” A tactile start edge and neutral adult description are available. Keep these exact lengths/positions out of practice.

**Mass setup:** Use two identical, equal-empty-mass, broad safe trays on a checked equal-arm balance. Put **three identical large safe blocks** in tray U and **six of the same blocks** in tray V. Verify *before class* that empty trays balance and V's loaded side goes down. If not, level/replace equipment and verify again; do not score a predicted answer from an unverified setup. Do not let the visible block count replace direct balance evidence. Ask “Which loaded tray is heavier, and what did you observe?” Allow child-directed adult placement or neutral audible/tactile balance description. Record any support and whether the child actually made a direct comparison.

**Exact key for verified set:** S is longer than R **after alignment**; V is heavier than U because its side of the checked balance goes down. Both claims are relative to their named pair. The length setup does not establish mass, material, capacity, strength or usefulness. A child who says “S looks longer” without aligning has not yet shown the direct-comparison method; offer a fresh different pair for recheck after teaching. If the actual balance behaves otherwise, fix it rather than marking a child wrong.

**Next move:** compares far ends with shifted starts → model common-start alignment on a *new* equal pair, then fresh unequal pair; length and width confused → name one attribute and use strips with same width; infers mass from apparent size → observe a preflighted balance with a larger empty/smaller loaded pair; balance relation reversed → change tray positions and ask which side moves down, then fresh load; correct results without reason → ask for the physical edge/balance indication rather than an adult preferred phrase; access method blocked → offer tactile/directing route and mark original observation **not observed**.

## Day 140 · fresh fictional image data

**Question to state before collection:** “Which paper story-board **picture tab** did these **eight invented visitors** choose: LEAF, BOOK or STAR?” Prepare **eight large image cards** from the teacher-only [original cut-card sheet](heldout-tab-cards.svg): **LEAF 4, BOOK 3, STAR 1**. Cut, mix in an envelope and never show the intact sheet before sorting. One image card = one invented recorded choice; inventory marks L1–L4/B1–B3/S1 go on backs only. Offer three empty labelled trays; read labels aloud if needed. Nonvisual cards: four leaf-outline, three book-ridge and one star-edge large safe shapes, with adult names read neutrally; one item once.

Ask child to collect/sort by **pictured tab shape**, show one card per response, identify most and fewest, compare BOOK with STAR by pairing/counting, state total if accessible, and say what one card represents. Ask what the set does **not** say about actual classmates or why the fictional visitors chose shapes. Preserve first group and relation before key feedback.

**Exact key:** LEAF **4**, BOOK **3**, STAR **1**, total **8**; LEAF most, STAR fewest; BOOK has **two more cards** than STAR (or simply more, if difference number is not the target). Group by picture tab, not card position/colour/back identifier. It is a teacher-made fictional dataset, so it cannot establish a real class choice, any absent respondent's choice or a reason. Accept one-to-one matching, counting, AAC/sign/voice or child-directed placement with prompt level logged.

**Next move:** wrong feature → contrast two large labelled shapes and sort a new four-card set; card lost/doubled → move each from unsorted pile once, then recount new set; count correct but comparison wrong → align matching pairs in separate rows, then new counts; reason or real-class claim invented → distinguish source of picture data from a respondent explanation and make a different fictional question; nine-card load too high → fresh two-category four/five-card diagnostic, retain Day 140 first attempt unchanged; secure sorting and bounded comparison → invite a different safe data question and display with a new fictional set, not an automatic real survey.

**Private record:** learner code/date ___; first R/S start alignment/result/reason ___; verified balance status and first U/V result/reason ___; Day 140 first collect/sort/count/compare/source boundary ___; adult label reading/access/help ___; revision ___; next lesson/recheck ___ . Follow school storage policy. No named peer, photograph, voice recording, home object or personal answer is required.

Original checks and keys © 2026 NeuroForgeIO Pty Ltd, [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). Credit, link and indicate changes.