# Separate teacher key · first responses before feedback

Use only after a child's first independent response to the corresponding [held case](ASSESSMENT.md). The case is new paper material; teaching F/G/H/J never supplies its particular positions or captions. Record exact access, neutral prompts and child's words. **There is no fixed local movement answer:** a real pad might follow any safe path, pause, stop, be interrupted or not be run. No paper case can certify a year standard, child progress, a real causal material property or actual safe investigation.

## Day 125 · Case AB

- **Source and literal evidence:** FICTIONAL PAPER CASE AB has an orange rectangle at LOW LEFT first, HIGH CENTRE second, LOW RIGHT third. The three pictures give that order and those positions. They contain no continuous line, interval observation, real material identification, date or witness.
- **Claim boundary:** The three positions could be joined in several ways; a zigzag path, two stops and cardboard cause are not shown. Even if “zigzag” seems plausible from low/high/low snapshots, a path between moments is not established. The printed claim cannot make material “cardboard” verified. A child may say “I don't know how it got between the frames.”
- **Good next source:** a locally approved adult-controlled large safe object's full path, dated, with start/end, clear view and raw pause/turn/stop notes, could answer a narrower question. The educator keeps approved conditions and access status; no child must handle an object.
- **Follow-up:** If the child reports the caption as fact, place F's three frames next to G's two possible paths after scoring and re-teach “positions shown / path unknown”. If positions are out of order, reread FIRST/SECOND/THIRD with fixed tabs. If the source is called actual, point to the FICTIONAL PAPER CASE label and contrast an actual dated record. Credit child-origin words or AAC/sign choices at the same target, not a preferred response mode.
- **Do not award:** an actual AC9SFU02 movement or AC9SFI02 investigation, a verified cardboard material, route, pauses, cause or real prediction from AB.

## Day 130 · Case AC

- **Source and literal evidence:** FICTIONAL PAPER CASE AC's imaginary adult account says the large green square was at START left, a paper screen hid the middle interval, and it was at FINISH right when the screen opened. The interval remains **NOT SEEN**. No real classroom date, real witness, actual source or prior prediction is supplied.
- **Claim boundary:** ROLLED THROUGH THE CENTRE and “because its material is soft” are separate unsupported printed assertions. The square could have been moved in more than one way during the hidden interval; the source does not select one. Softness is unverified and is not a causal explanation established by endpoints.
- **Honest private share:** “In this invented account, the green square is reported at START and later FINISH. The middle was hidden. I cannot tell whether it rolled or why.” A next genuine source requires adult approval, local date, known safe large object/table, direct clear access to the full interval or exactly attributed ADULT REPORT, raw action/positions, and explicit interruption note if the view fails.
- **Follow-up:** If the child fills the hidden interval, cover the middle of a blank route mat and practice keeping it blank. If a child treats ADULT REPORT as their own direct sense, relabel who observed and who heard. If a prediction is invented after the endpoint, practise “NO PRIOR PREDICTION” before another scenario. Accept “cannot compare” without penalising the child for missing source data.
- **Do not award:** actual AC9SFI02 investigation, a real movement/path, real material cause, prior-prediction comparison, child improvement or listener response from this invented account. An imaginary observer in a fictional case is **not the child's own direct sense**.

**Interpretation and next action:** At Foundation, source-bounded reasoning can be shown through speech, sign, AAC, pointing, drawing or adult-directed words. Record the response and support instead of a global “understands motion” label. If an approved genuine investigation also happened, assess that separately from these cases using the actual dated log. If not, record **REAL OBSERVATION NOT AVAILABLE** and plan an accessible local opportunity only if safe. Original key © 2026 NeuroForgeIO Pty Ltd, SubjectNest, [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/).