# Foundation / Prep Science · ten 25-minute scripts, Days 91–100

Read [adult gate and complete original Models M/N](MATERIALS.md) first. Each lesson totals **3 + 4 + 5 + 7 + 4 + 2 = 25 minutes**. Use [three access-equivalent routes](LEARNER-CARDS.md) flexibly and [two worked swaps/day](OPTIONAL-PRACTICE.md) when useful. The carton apparatus is adult-only, pretested, locally approved or absent. The paper fallback is **FICTIONAL MODEL** throughout. With no safe dated actual run, write **REAL OBSERVATION NOT AVAILABLE**. A neutral spoken **ADULT REPORT** is attributed, not the child's own direct sensory observation. Never turn a fictional stop into a classroom result. The [fresh Days 95/100 checks](ASSESSMENT.md) stay teacher-held until each first independent response. Family invitations carry no assessment burden.

## Week 19 · Day 91 — What exactly would our test ask?

**Goal:** pose a clear movement question and record a prediction *before* any genuine result. **Proposed codes:** AC9SFU02, AC9SFH01, AC9SFI01. **Prepare:** [question mat](question-predict.svg), complete M text, approved setup only if already cleared, Card 91.

1. **Name source · 3 min.** Mark M FICTIONAL MODEL. State separately whether an approved actual carton trial is possible; if not, REAL OBSERVATION NOT AVAILABLE.
2. **Notice · 4 min.** Read M's setup paragraph without results. Ask what changes when the same carton rests on broad rather than narrow face.
3. **Model · 5 min.** Teacher asks “Does the carton reach the same ordered mark on each face?” and writes a possible prior prediction with time/order, not a fact.
4. **Child try · 7 min.** Card 91 routes make/select one testable question and prior prediction. Adult records exact wording or AAC selection before any result is shown.
5. **Check · 4 min.** Offer “We can make a prediction after reading the stops.” Child distinguishes an advance prediction from an after-result guess.
6. **Exit · 2 min.** Save question/prediction separately from later observation; no outcome inferred.

## Week 19 · Day 92 — Which things should stay steady?

**Goal:** identify one planned change and multiple steady conditions in a fair comparison. **Proposed codes:** AC9SFU02, AC9SFI01, AC9SFI05. **Prepare:** [setup map](setup-change.svg), M first paragraph, broad fixed labels, Card 92.

1. **Recall · 3 min.** Revisit yesterday's question and label it M or actual plan, never an actual result.
2. **Read · 4 min.** Read “same carton, tray, start line, tilt, no-push release”; name face placement as the planned change.
3. **Model · 5 min.** Teacher deliberately moves a paper START label, then says that would be another change and spoil a clean comparison.
4. **Child try · 7 min.** Card 92 routes place/say one-change and three-steady labels; teacher asks which changed condition must be corrected before a new trial.
5. **Check · 4 min.** Ask if two different cartons plus two different tilts could isolate face placement. Child says why not, without needing technical vocabulary.
6. **Exit · 2 min.** Retain child-made same/change plan as planning evidence only.

## Week 19 · Day 93 — What if a run gets interrupted?

**Goal:** keep a raw trial record and label an interrupted run without silently treating it as a valid stop. **Proposed codes:** AC9SFI02, AC9SFI03, AC9SFI05. **Prepare:** M's fifth entry, [four-run sheet](four-run.svg), optional adult-pretested apparatus, Card 93.

1. **Safety gate · 3 min.** Adult verifies tray, lip, carton, supervision and quiet access, or selects M only. Children may observe or use description without touching.
2. **Read · 4 min.** Read M fifth entry: carton tipped and touched the rim. Say it is invented and marked INTERRUPTED.
3. **Model · 5 min.** Teacher preserves the raw entry but excludes it from valid-stop comparison; model a cautious reason, not “it would have reached STOP.”
4. **Child try · 7 min.** Card 93 routes classify VALID versus INTERRUPTED from a labelled paper entry; if genuine safe rehearsal occurs, record date, apparatus and what actually happened separately.
5. **Check · 4 min.** Offer “Cross out the tipped run so our record looks neat.” Child keeps it, labels it, and asks whether a safe repeat is possible.
6. **Exit · 2 min.** Save raw/unrevised record; a rehearsal alone is not a four-run comparison.

## Week 19 · Day 94 — What did the valid tries actually show?

**Goal:** make a bounded comparison of four raw attempts and separate genuine evidence from M. **Proposed codes:** AC9SFU02, AC9SFI02, AC9SFI03, AC9SFI04. **Prepare:** four-run sheet, claim mat, school-approved apparatus *only if safe*, otherwise M, Card 94.

1. **Choose source · 3 min.** Write ACTUAL with local date/approval/equipment if an investigation is genuinely run; otherwise M FICTIONAL MODEL and REAL OBSERVATION NOT AVAILABLE.
2. **Read prediction · 4 min.** Retrieve the actual advance prediction if one exists. If not, mark NO PRIOR PREDICTION; do not manufacture one.
3. **Model raw record · 5 min.** Adult models one labelled attempt without a push, or reads M's four original valid entries exactly; retain interruptions.
4. **Child try · 7 min.** Under approved conditions adult conducts/records remaining brief attempts or reads M in order. Card 94 routes mark face, position and interruption for each; if too few valid attempts, say insufficient comparison.
5. **Check · 4 min.** Compare exact valid stops to the advance prediction with guidance. “In these tries …” is allowed; “this face always slides farther” is not.
6. **Exit · 2 min.** Save raw dated genuine record or labelled M and the child's exact bounded claim.

## Week 19 · Day 95 — Fresh fair-test check

**Goal:** independently detect a second changed condition in a new paper trial and state what a real test would still need. **Proposed codes:** AC9SFU02, AC9SFI01, AC9SFI03, AC9SFI04. **Prepare:** [teacher-held Case R](ASSESSMENT.md), blank [setup map](setup-change.svg), Card 95 only after response.

1. **Preflight · 3 min.** Confirm Case R was unseen; prepare full read-aloud or secured broad tactile equivalent, neutral scribe and private first-response record.
2. **Present · 4 min.** Read complete new FICTIONAL case once without giving the fair-test answer.
3. **Think · 5 min.** Give quiet independent response time. A neutral reread is recorded, never an answer cue.
4. **First independent response · 7 min.** Child names planned change, another changed condition, what can and cannot be concluded, and a safer fairer actual setup; preserve exact wording/support before feedback.
5. **Source bound · 4 min.** Ask what an approved, dated real run would require. Paper-only answer remains REAL OBSERVATION NOT AVAILABLE.
6. **Close · 2 min.** Use [separate key](TEACHER-KEY.md) for a next teaching move; Card 95 follows later.

## Week 20 · Day 96 — Which Science sample has a real source?

**Goal:** select one genuine dated Science record or one explicitly invented sample and name what it can show. **Proposed codes:** AC9SFH01, AC9SFI03, AC9SFI05. **Prepare:** [source portfolio](source-portfolio.svg), verified child-owned consented private samples if available, M/N paper strips, Card 96.

1. **Privacy/source gate · 3 min.** Offer only consented private child-owned school samples with actual source/date; otherwise M/N. No public name/photo/location or family record.
2. **Read · 4 min.** Read N strips fully. Strip 1 has invented attempts; strip 2 has an unsupported “always” claim.
3. **Model · 5 min.** Teacher chooses M as model evidence for a *paper-record* explanation and marks unsupported N strip 2 as no result.
4. **Child try · 7 min.** Card 96 routes select a verified actual source or M; label DIRECT / PHOTO / ADULT REPORT / MODEL and source/date.
5. **Check · 4 min.** Ask whether an unlabelled picture or attractive sentence proves a real result. Child names a missing source or raw observation.
6. **Exit · 2 min.** Save private sample choice or model choice, without implying child growth.

## Week 20 · Day 97 — Can a claim point back to its marks?

**Goal:** connect one bounded statement to raw stops, including interruptions and uncertainty. **Proposed codes:** AC9SFU02, AC9SFI03, AC9SFI04. **Prepare:** actual dated run if present or M, [claim-limit](claim-limit.svg), Card 97.

1. **Source label · 3 min.** Display ACTUAL/date or M MODEL at every stage. Do not substitute M for an actual mixed result.
2. **Re-read · 4 min.** Read all valid positions and any interrupted entry aloud in original order.
3. **Model · 5 min.** Say of M, “In this invented set, narrow-face tries both stopped at MIDDLE; broad-face stops differed. This does not tell what a real carton will do.”
4. **Child try · 7 min.** Card 97 routes point/say/scribe one claim with the supporting exact entries and one unknown. If actual attempts are too few or mixed, “not clear yet” is valid.
5. **Check · 4 min.** Offer unsupported N strip 2. Child rejects “always” because four pretend entries cannot carry it.
6. **Exit · 2 min.** Keep child's actual wording, source, prompt and uncertainty.

## Week 20 · Day 98 — Did I see it, or did someone tell me?

**Goal:** distinguish direct observation, attributed adult report and fictional paper data in a handover. **Proposed codes:** AC9SFH01, AC9SFI02, AC9SFI05. **Prepare:** [source portfolio](source-portfolio.svg), verified actual source if any, M, neutral description line, Card 98.

1. **Name access · 3 min.** Ask what source each child actually accessed; never infer from a response mode.
2. **Read · 4 min.** Teacher says a *model-only* neutral line: “M says broad try 1 stopped at START.” This is a report about fiction, not a child's direct observation.
3. **Model · 5 min.** Contrast “I saw the actual dated carton stop” (only when true), “the adult told me …” (ADULT REPORT), and “the model sheet says …”.
4. **Child try · 7 min.** Card 98 routes label one actual/source sample and rehearse an exact evidence sentence with correct attribution; if no real investigation, say REAL OBSERVATION NOT AVAILABLE.
5. **Check · 4 min.** Ask whether hearing an adult describe a test gives the child direct sight. Child attributes the adult and can still reason from the report.
6. **Exit · 2 min.** Retain access/source label and support, not an assumed modality identity.

## Week 20 · Day 99 — Teach someone one honest Science idea

**Goal:** privately explain question, source-bounded evidence, limit and next safe check to a chosen audience. **Proposed codes:** AC9SFH01, AC9SFI04, AC9SFI05. **Prepare:** [private handover](handover.svg), actual consented source or M/N, Card 99.

1. **Audience choice · 3 min.** Offer teacher, trusted partner or private adult-held recording in approved system; no public sharing requirement.
2. **Order · 4 min.** Place SOURCE, QUESTION, RAW EVIDENCE, CLAIM, UNKNOWN in sequence; keep real/model label on every card.
3. **Model · 5 min.** Adult gives a complete M explanation and labels its status: pretend narrow results at MIDDLE twice, unknown real outcome, next approved run.
4. **Child share · 7 min.** Card 99 routes communicate their chosen bounded explanation; adult records exact independently selected content and prompt.
5. **Feedback · 4 min.** Log an actual listener question/response only if it occurs; otherwise **NO ACTUAL LISTENER RESPONSE**. Revise from source review if useful.
6. **Exit · 2 min.** Keep private record with one next safe question, not a progress score.

## Week 20 · Day 100 — Fresh evidence-handover check

**Goal:** independently reject an unsupported conclusion from a novel paper record and plan what genuine source could test it. **Proposed codes:** AC9SFH01, AC9SFI01, AC9SFI03, AC9SFI04, AC9SFI05. **Prepare:** [teacher-held Case S](ASSESSMENT.md), blank [handover](handover.svg), Card 100 after first response.

1. **Preflight · 3 min.** Ensure Case S was unseen; offer exact full oral/broad secured tactile equivalent and private recording route.
2. **Present · 4 min.** Give complete FICTIONAL case with all raw marks, source labels and claim.
3. **Think · 5 min.** Independent quiet time; neutral reread only, with prompt noted.
4. **First independent response · 7 min.** Child identifies source, supported statement, unsupported claim, prior-prediction limit and next approved actual evidence; capture first wording/access before key.
5. **Bound · 4 min.** Ask what could be said to a private listener without claiming an actual class result. If no local genuine source, REAL OBSERVATION NOT AVAILABLE.
6. **Close · 2 min.** Use [separate key](TEACHER-KEY.md); this check does not certify year/term achievement.