# Teacher-targeted worked guide and next teaching moves

This guide is **publicly accessible** with the pack. The Days 15/20 prompts are fresh against the lesson practice, **not secure exams**. If a learner may have seen either prompt or this answer, replace it locally for any consequential evidence decision. Use each criterion as `0` absent/incorrect, `1` developing or supported, `2` secure **in this sample** with source evidence. Record interpretive prompting apart from reading/AAC/typing/Braille access; retain first work and changes. Do not sum the eight points into an English grade or infer independent print reading from an adult read-aloud.

## Daily observation and response guide

| Day | Evidence to notice | If fragile, teach next |
| ---: | --- | --- |
| 11 | The unit is **instructed page turns per scripted route**; five complete pairs A30/B22, not speed or person experience. Route 4 ties; the sixth B endpoint is missing. | Put a COUNTED and NOT TESTED card beside the source. Ask “of how many complete pairs?” before any outcome word. |
| 12 | A repeatable paper explanation names same six prompts, two drawn layouts, arrow trace, page-turn count and the incomplete sixth B route. | Reorder question → construction → unit → missing decision; replace vague “tried” with “traced on paper”. |
| 13 | Text B's “six”, “every time”, “faster” and “order today” exceed Text A; a fair title names five paper pairs and instructed turns. | Attach each notice phrase to a source line or UNKNOWN card; bring limit into the first sentence. |
| 14 | A defensible ethical reading cites at least two of Mara's notice, Ivo's question, visible blank card, heading moved up; an alternative acknowledges ambition to run a real test or caution. | Read the final heading action aloud; ask what changed for a reader although the paper counts did not. |
| 15 | Use fresh Check A criteria A1–A4 below. | Recheck with a new locally authored missing-record source after the relevant next move. |
| 16 | F's 21/30=70% concerns named routes; G's 22/36≈61.1% includes all tags. F's exclusion note is late and its title lacks a group; G's full total is in its opening. | Highlight the reference group directly under each headline, then locate the end note without prompting interpretation. |
| 17 | The six route-unreadable tags have known statuses, 1 same-day/5 later. “Although... therefore...” keeps the all-tag total visible without assigning cause. | Separate ROUTE and STATUS fields with two tactile/printed cards; then add the linking clause. |
| 18 | Learner explains how an evaluative opening foregrounds a positive subset or labels records unfairly; revision names group and value before inference. | Compare openings read alone; ask what group the reader would assume and what the table can actually say. |
| 19 | Two-part report starts with 22/36 all-tag figure, treats 21/30 as a scoped named-route comparison, mentions six unreadable routes and avoids causal or real-service claims. | Give source, scope, figure, limit headings without a supplied conclusion; have learner reorder own sentences. |
| 20 | Use fresh Check B criteria B1–B4 below. | Recheck with a new locally authored categorical register and two differently framed report leads. |

## Check A Day 15 worked guide

**Source arithmetic and boundary:** Only outlines 1, 2, 3 and 5 have a complete R/S pair. R = `4+5+3+4=16` panels; S = `3+4+3+2=12`. For those four outlines, averages are 4 and 3 panels, a difference of **one diagram panel on average**. Individual paired differences are 1, 1, 0 and 2 panels, so “every time” is false; outline 3 ties. Outline 4 has R6 and **no S count**: treating its blank as zero would invent a measurement. No lantern was built and no people used the instructions, so “builds faster for everyone” and “tested” real use are unsupported.

**Example corrected lead:** “For four complete pairs of **drawn** paper-lantern instructions, set S uses one fewer diagram panel on average than R (3 versus 4 panels per outline). One pair ties, the fifth pair lacks an S finish, and no build time or reader access was tested. A next step would be to complete the missing drawing and separately plan an accessible real-use trial with consent if one is needed.” Equivalent precise and source-bound wording is valid; no student must endorse a real trial.

| Criterion | 0 · absent/incorrect | 1 · developing/with interpretive prompt | 2 · secure in this sample | Immediate next move for 0/1 |
| --- | --- | --- | --- | --- |
| A1 · question/method/unit | Calls counts speed or claims people built lanterns. | Names diagram panels but omits matched outlines or paper-only construction. | Explains same five drawn outlines, R/S panel counts and no build/person test. | Sort SOURCE METHOD and REAL USE statements; rewrite the counted unit. |
| A2 · complete-pair comparison | Uses five pairs, treats blank S as zero or gives wrong total. | Gets 16/12 or 4/3 but omits four-pair scope or tie. | Shows four complete pairs, correct totals/averages and tie, with an exact row. | Cross out incomplete row, total four complete rows, then inspect outline 3. |
| A3 · notice audit | Accepts universal speed/test claim. | Challenges one phrase but does not tie it to source evidence. | Explains at least two overreaches with missing pair/tie/unit/no-use evidence. | Put ALL FIVE, EVERY TIME, FASTER and TESTED above matching source cells or UNKNOWN. |
| A4 · revised lead/limit | Repeats the overclaim or has no source condition. | Makes a cautious sentence but hides the paper/complete-pair boundary or next question. | Gives a clear source-bound lead, visible untested limit and relevant next question. | Order question → paper result → missing record → untested use; reread first sentence alone. |

**Criterion response:** A1 → method/unit language; A2 → missing record and denominator; A3 → source-to-claim challenge; A4 → report order and qualification. A fluent oral/AAC response with exact transcript can show the same reasoning; public speaking confidence is not the construct.

## Check B Day 20 worked guide

**Source arithmetic and boundary:** East/West named trays have `16+14=30` complete of `20+20=40` cards: **75%**. With unreadable-tray cards included, `30+2=32` complete of `40+10=50` total: **64%**. The ten unreadable-tray cards have known statuses **2 complete, 8 pending**. H's first sentence omits the reference group and can be read as all 50; its end note discloses the omitted category only after a confident title and “clearly working” conclusion. I makes the full denominator visible immediately. Neither report can infer cause, real system performance or what people experience.

**Example corrected lead:** “Of the **40 invented cards with East or West tray labels**, 30 are marked complete (75%). Across all 50 invented cards, including ten with an unreadable tray label, 32 are marked complete (64%). These status counts do not show what caused completion or whether any real system works.” A learner can prefer a different order if both scopes and the structural effect are clear.

| Criterion | 0 · absent/incorrect | 1 · developing/with interpretive prompt | 2 · secure in this sample | Immediate next move for 0/1 |
| --- | --- | --- | --- | --- |
| B1 · denominator/category | Treats 75% as all 50 or drops unreadable-tray statuses. | Gets one fraction or both without explaining ten-card difference. | Locates 30/40 and 32/50, includes unreadable tray 2/8 and keeps status separate from route. | Build two labelled subset/all cards with 40 and 50 tokens, then attach counts. |
| B2 · report structure/evaluation | Says only “H is biased” or invents motive. | Notices late end note or strong wording but no likely reader effect. | Explains how H's title/lead/end note or I's early denominator shapes a reasonable reader's inference, without claiming intent. | Read each first sentence alone; ask “of how many?” before looking at notes. |
| B3 · revision | Repeats “three quarters of the cards” as all cards. | Names 40 in revised lead but omits 50-card follow-up or has unclear group. | Gives accurate 30/40 named-tray lead and 32/50 all-card follow-up, with readable scope. | Move the denominator into the opening clause, then add a separate all-card sentence. |
| B4 · limit | Treats status as cause, real performance or people data. | Says “fictional” without specifying what cannot be inferred. | States this is a constructed status register, no cause/real performance, and tray-unreadable does not erase status. | Separate OBSERVED-IN-MODEL, NOT MEASURED and REAL-WORLD UNKNOWN. |

**Criterion response:** B1 → reference group and field distinction; B2 → reverse outline and reader effect; B3 → front-loaded scope; B4 → causal and real-world limits. If prior access to these public answers is plausible, do not reuse the printed prompt to claim independent transfer.

## Optional extra quick notes

Day 11 invented board-game counts must remain drawn-path counts. Day 12 method language avoids “tested with readers”. Day 13 “all six” must not swallow the missing route. Day 14 an alternative interpretation needs exact Text D evidence. Day 16 gallery example deliberately has 6 named and 2 unreadable routes: a percentage of named records would use 6, an all-record count 8. Day 17 route and status are separate fields. Day 18 a neutral label removes an invented fault judgement. Day 19 the audio or visual route needs a complete transcript/caption. Days 15 and 20 extras happen **after** each fresh check is collected.

**Original resource rights:** © NeuroForgeIO Pty Ltd 2026, SubjectNest, [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). Credit author, source, licence and changes.