# Teacher only · open keys, descriptive rubric and next moves

Keep this and model-data.json away from the first learner C/D capture. Record exact source access, first/later response and help. Source descriptions can be read; do not supply a question, change, answer or ideal height as the learner's first thinking. An unfinished component is **not yet observed**: offer another short sitting; **no completion-speed score**. No numerical projection, all-bank choice, full checklist or finished artwork quota.

## C — independently enumerated valid examples

Exact source: point `(0,0)`, bar `x=2`, height2 centred `y=0`, screen6; all arbitrary ideal-model units, no actual trial/permission supplied. General ideal span for this centred bar is `bar height × screen distance / bar distance`; it is a teacher geometry consistency relation, not Year 5 learner ratio assessment. Initial ideal span6, but do not fill that answer on the learner aid or copy it into an actual-data record.

Accept an own question/prediction/plan where its chosen variable, comparison and evidence are coherent. Examples, independently checked:

1. **Change only bar position:** bar2→1 with source0/screen6/size/orientation retained. Predict a larger ideal span (12 versus6), or offer a qualitative reason about the blocked paths. An actual extended lamp may differ; no actual result supplied.
2. **Change only bar position another way:** bar2→3, otherwise same. Predict a smaller ideal span (4 versus6). Different valid positions between source and screen are welcome; no particular one required.
3. **Change only screen position:** screen6→8, bar2/source0 retained. Predict a larger ideal span (8 versus6). Own safe actual distances would be chosen locally, never copied as room instructions.
4. **Different investigable feature:** choose blocker size/orientation or a suitable screen-material/clarity question, identifying a useful observation and appropriate retained conditions. Such a plan need not seek an exact model span; a justified uncertainty can be appropriate.
5. **Own/rejected source:** choose an appropriate attributed alternative or own model and coherent light question. A chosen creative purpose/rejection is valid participation even where investigable reasoning is not yet observed. Do not automatically award I01/I02 from choice alone.

A partial question, prediction/reason or plan is captured first. Later components may develop it. If multiple conditions change, name that as exploration and invite a clearer comparison; do not call it a one-variable test. No apparatus view, equipment skill, human-source breadth or real audience effect comes from this paper brief.

## D — independently checked open repairs

Exact before model source0/bar3/screen9, proposed after source0/bar2/screen6, same bar height2/orientation, arbitrary ideal units. Both bar and screen positions change. Ideal spans are `2×9/3=6` and `2×6/2=6`, not a larger model span. This does **not** mean real lamps always make the same shadow or that an actual test occurred. The proposed cause claim is unsupported as a single-variable finding, and no actual measured result/permission is supplied.

Valid first responses include identifying the two changes, a coherent repair, an investigable own question or insufficient evidence. Examples:

1. **Retain screen9:** compare bar3→2 with source0/screen9/size/orientation fixed. Ideal spans6→9 (qualitatively larger). Actual trial separately reviewed/recorded.
2. **Retain bar3:** compare screen9→6 with source0/bar3/size/orientation fixed. Ideal spans6→4 (qualitatively smaller).
3. **Retain proposed screen6:** compare bar3→2 with source0/screen6/size/orientation fixed. Ideal spans4→6. This is a new comparison plan, not the brief's before result.
4. **Retain proposed bar2:** compare screen9→6 with source0/bar2/size/orientation fixed. Ideal spans9→6. Do not invent these as earlier observations.
5. **Own coherent comparison:** change another suitable single variable, define what is retained/observed, or propose a different investigable light question. There is no mandatory left/right movement, fixed physical range or unique repair.
6. **Evidence limit:** “No actual trial yet,” “two things changed,” or a reasoned need for comparable observations can be a meaningful component. A vague refusal/choice alone is valid participation with inquiry reasoning open, not automatically incorrect science.

After first capture, invite a reason/prediction/evidence need or later trial if useful. Use the original first record alongside later evidence, not a backfilled answer.

## Descriptive rubric · available task evidence, not a grade

| Dimension | Not yet observed / next invitation | Developing evidence / useful move | Available coherent evidence / useful move |
| --- | --- | --- | --- |
| Question/prediction | No available investigable question or reason yet; offer one source fact at a time, invite an own question. | Identifies a change or expected effect with partly specified comparison/reason; ask what might be noticed. | Own coherent question/prediction with a reason and relevant comparison; explore another appropriate question without a quota. |
| Comparison/model planning | A plan/source choice is available but retained conditions/evidence need is unclear; ask one useful “what stays?” | Some conditions named or a changed condition recognised; clarify one remaining relevant condition. | Coherent comparison/model limits and evidence need; consider an actual locally approved test if wanted. |
| Actual inquiry roles | Paper/director/source-only response; actual operation/observation/measurement not supplied. Preserve the participation route. | Actual event available with roles/help/edge/unit partly recorded; clarify provenance/precision. | Actual operations/observations/measurements attributed at their demonstrated scope; consider a useful repeat, not automatic full I02/I03 award. |
| Representation/evidence | Model prediction/actual record source unclear; label source/type and leave missing results blank. | Useful representation with some conditions/provenance, tentative relation; seek missing information. | Clear appropriate available-data/model representation, justified relation/uncertainty, no fabricated results; consider another method/source where available. |
| Communication/purpose | An intended purpose exists but learner meaning is not available yet; invite one component in familiar mode. | Communicates a choice/reason with help or partial audience/purpose relation; develop one useful detail. | Actual learner communication fits the chosen purpose/intended recipient with help named; actual audience response remains conditional. |

These descriptors do not assess speech, motor range, posture, neatness, compliance, speed, reading independence or child worth. Another sitting is an evidence opportunity, not a penalty. A supported response can be substantive learner reasoning; name the support precisely.

## Ordinary-day next moves

- Source versus screen confused: encounter the actual emitting source where available, or identify roles in the named model. Do not infer a real device state from a torch icon.
- Ideal numbers appear in actual-data rows: preserve the original response, label ideal prediction separately and leave actual rows missing until available.
- Two conditions changed: retain the exploratory observation/plan, choose one useful controlled comparison; no invented repeat.
- Shadow edge unclear: agree an approximate edge/qualitative outcome, record uncertainty; do not manufacture instrument precision.
- Model fails to explain a real view: keep the view, ask about extended source/ambient light/alignment as possibilities, not established causes.
- No other investigator or willing reader: keep those components open; own source/model/communication work remains useful.

This is formative, partial Year 5 evidence. Ten authored starts and two checks do not certify U03, all inquiry rows, the achievement standard, a whole year or a jurisdictional programme.