# Teacher held · fresh formative checks, Days 15 and 20

Hold this page and the [separate key](TEACHER-KEY.md) from learners until a **first independent** response is saved. Offer pointing, speech, sign, AAC, drawing, tactile broad cards, exact adult read-aloud and no-touch participation; record the support and access route. Do not cue the answer. Date/local source/action/result may be written by an adult verbatim. For an actual trial, complete the [safety and provenance preflight](MATERIALS.md) first. No approved trial means **REAL OBSERVATION NOT AVAILABLE** and a paper-only reasoning check. The models below are original and have not appeared in [learner cards](LEARNER-CARDS.md) or [worked swaps](OPTIONAL-PRACTICE.md). Neither check certifies a whole description, standard or year.

## Day 15 · unseen source and path

**ACTUAL LOCAL SOURCE option:** Teacher selects a new, clean, intact, broad, light, soft school-approved object not used in Days 11–14; a broad padded rectangular prop is a candidate only if locally approved. Record exact object, verified material or UNKNOWN, date, surface, START/CATCH, release action, adult handler and access route. The adult gently releases it once along a clear flat lane with a catch after the child's prediction; actual movement may be slide, turn, stay or something else. No result is prewritten. Do not identify the object as the one in the model below.

**PAPER MODEL ONLY fallback, fresh and invented:** A wide horizontal rectangular lane. START is at centre; CATCH LEFT is at the left edge and CATCH RIGHT at the right edge. A large **square outline** sits at START. One arrow points left and stops halfway to CATCH LEFT; a second arrow points right and stops close to CATCH RIGHT. There is no object, release, observer, material, time, scale, actual path or local result. Exact read-aloud/tactile: “START in the centre; a large square outline at START; one arrow goes left to halfway; another goes right to near the right catch.” Use a broad fixed strip and two large arrow cards if tactile, labelled INVENTED MODEL.

Ask in this order, using a blank [motion-question](motion-question.svg) and [motion-log](motion-log.svg):

1. “Is your source a real object here, an adult report about it, or an invented board? What source and date or model-use date shall we write?”
2. “Ask one question about how the *real approved object* might move. Give your prediction **before** any release.” If only paper, ask a question a future safe real try could answer; leave actual result NOT TESTED.
3. With actual option, adult performs approved release and asks “What did you directly notice, or what did the adult report?” With model fallback, ask “What marks and directions does the board show? Does it prove two objects moved?”
4. “What is still unknown from your source? May we compare your prediction with an actual result, or must we mark NOT TESTED?”

**Capture:** first independent question/prediction, exact observation or model description, source label, unknown, prompt/access and whether an actual result exists. Release Day 15 practice card only afterward.

## Day 20 · unseen comparison and factor limit

**ACTUAL LOCAL SOURCE option:** Before the check, adult finds two new broad, light, intact, soft, school-approved objects with **same verified material and approximately same shape but different verified sizes**; if no sufficiently comparable pair can be sourced, do not use this option. Record exact identity, material-verification source, size description, local date, lane surface, same marked START/CATCH, adult gentle release method, catch and access route. Note any unmatched weight, finish or stiffness. Have child predict A/B separately before adult conducts one safe try with A and one with B. Write results exactly; if any condition changed, record it. No result, direction or size effect is prescribed.

**PAPER MODEL ONLY fallback, fresh original fictional two-record source:** The two cards are marked **INVENTED; NO CLASSROOM TRIAL**. Record A: “A broad soft **ring-shaped** prop of unknown material starts at a marked centre. A fictional observer wrote STOP AT CENTRE.” Record B: “A broad soft **flat oval** prop of unknown material starts at the same marked centre. A fictional observer wrote STOP TO THE LEFT.” The cards give no dimensions, release method, real lane surface, date of observation or actual object. They vary shown shape and perhaps more; the differing invented stop labels cannot establish that shape caused movement. Exact read-aloud/tactile: read A fully before B; use two broad source cards and large CENTRE/LEFT endpoint labels, all labelled invented.

Ask in this order with a blank [compare-factors mat](compare-factors.svg):

1. “Which sources are you using? Are they actual local trials or invented records? What is known and unknown about each?”
2. “Before any *actual* release, what do you predict for A and B?” For fallback, a child may say what they would predict in a future safe trial, but the supplied fictional outcomes are not actual results and must not be called a prediction check.
3. For actual option, ask “What did each approved object actually do here? Did your predictions match these observations?” For fallback, ask “What stop words do the two invented cards say? Can they prove a local result?”
4. “What was the same, what was different, and can you say size or shape caused the difference? What further safe check or control would you need?”

**Capture:** exact first response, source distinction, A/B comparison, causal limit, prompt/access and next safe check. If a matched real pair is unavailable, use fallback and do not assign actual movement or safe-investigation evidence from the model.

Original assessment prompts and fictional sources © 2026 NeuroForgeIO Pty Ltd, [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/).