Use this apart from the learner page when facilitating. The new-case checks and this key are publicly accessible; introduce checks on Days 15/20, and substitute a locally unseen equivalent when prior access matters. This is a formative guide, not a secure or validated rubric. Accept a different defensible interpretation when the learner cites an exact word/detail and does not claim an unsupported real outcome. Record I independent, A neutral access support, P content prompt or N not observed for each construct. Do not infer independent reading from a neutral read-aloud; do not mistake an access choice for a content error.
Daily exits and immediate next teaching
| Day | Expected evidence | If not yet secure |
|---|---|---|
| 11 | because fits table blocking arrow; second sign's effect unobserved. later only orders events, therefore cannot supply missing evidence. |
Match each connective to a relation card: time, reason, contrast, supported result. |
| 12 | Text C's it can be read ambiguously; a clear repair is “The table blocked the east-door view of the arrow.” In the exit, name table as blocking the view rather than leave it open. |
Put a noun card above every pronoun; if two nouns fit, replace pronoun. |
| 13 | relocation names the move; after actor/action is known, “the move” can summarise it. Text A does not prove confusion was caused, improvement occurred, or confidence returned. |
Ask “Who did what?” then “What does the source actually show?” |
| 14 | Text A records two visitor questions; it has no confidence feedback/measure. Clear note attributes crew actions and explicitly names unknown later effect. | Split one elegant but overconfident sentence into recorded action and unanswered question. |
| 15 | Fresh key below. | Diagnose referent, nominalisation, source limit and reader-facing revision separately. |
| 16 | “Hesitation” or “uncertainty” can be a supported interpretation of Mica's crossed-out word and silence; no exact emotion/diagnosis is stated. | Underline action first; add “suggests” to inferred state. |
| 17 | A safe report line: “The crew added a second arrow; the log does not show whether it helped.” Confidence returned is absent from log. |
Compare a logged noun (“addition”) with unsupported outcome (“confidence”). |
| 18 | Seven balls counted after practice versus eight before; borrowing/cause not recorded. Therefore someone stole one is unsupported. |
Replace causal connective with afterward; keep cause unknown. |
| 19 | Any purposeful old/new edit that clarifies referent or changes a reader's likely experience, with concrete action before abstract noun and no fixed assigned emotion. | Ask an absent reader what “it/that” names; try one literal action before a mood noun. |
| 20 | Fresh key below. | Separate report evidence, fiction inference, cohesion and audience edit. |
All thirty route keys and feedback
| Day | A | B | C | Next move |
|---|---|---|---|---|
| 11 | Text A order 9:00 place, 9:20 move, by 9:35 two ask; because table blocked view, later questions. |
Music card: host posted new order; one performer asked if ending time changed; final programme absent. | Sports card: 8 before/7 after; theft not documented. “After practice, seven were counted; the reason for the difference is unrecorded.” | Request a source line for every result connective. |
| 12 | they→crew; those two questions→two visitor direction questions; replace ambiguous C clause with “because the display table blocked the view of the east-door arrow”. |
A safe theatre rewrite: “The stage manager moved the rehearsal to Room C. Two performers asked for the new start time.” The original it obscures the question's subject. |
B's this decision could name the first move or second arrow; choose “the second-arrow addition” only if that is the intended link, without claiming success. |
Put the full noun after a reference word once, then test reader. |
| 13 | relocation→crew moved east-door arrow; useful after the action is named, but “caused confusion” is an extra causal claim. |
“The organiser shifted the paper labels. After the label move, a visitor asked about a box; the later effect was not recorded.” | One fewer foam ball was counted; neither borrowing nor theft was recorded. “The difference in count needs checking.” | Restore actor/action before summarising. |
| 14 | Accept a 3–4-sentence note with invented source status, crew move, two questions, second arrow, no known later effect. | Theatre script states room move and two start-time questions; start time/attendance unknown. | C's vague it can name the blocked arrow view; B's “confidence returned” must be removed/qualified as unsupported. |
Keep source status and limit visible to absent reader. |
| 15 | After check theatre the move = stage manager changed Room A to C. |
Seed-swap organiser shifted labels after wind lifted one; later outcome unknown. | Music card gives a new order and one question, not proof of improvement. | Do not teach with the held-out demo log before a recheck. |
| 16 | Two actions such as drew routes/crossed out left; noun “hesitation” is interpretation with cue. |
Speech says “may suggest uncertainty” and cites a line; avoids assigned diagnosis. | certainty conflicts with crossed-out direction and “ask at desk”; “hesitation” is more plausible, though not only reading. |
Quote action and mark interpretation. |
| 17 | Addition=second arrow documented; improvement=unsupported valuation; hesitation=fiction interpretation. | Report can establish action/unknown effect; fiction allows supported mood reading. | “Crew added a second arrow; effect not logged.” “Mica's correction and silence may suggest uncertainty” rather than fact of fear. | Ask if a noun reports an observation, an evaluation or an inference. |
| 18 | Sports: eight counted before, seven after, cause not recorded; afterward is safe. |
Seed labels shifted, visitor later asks; “the label move” clear; no observed resolution. | Theatre room move and two questions do not establish any announcement or universal attendance. | Use relation word for known connection only. |
| 19 | Original 55–70-word scene with concrete action, clear reference, one justified abstract noun, saved edit and reader effect. | Same literary evidence in an audio/AAC script; no recording required. | Replace vague it/them with anchored nouns and show an action before inferred “worry”; many creative solutions valid. |
Check clarity then style; never grade a particular emotion as mandatory. |
| 20 | After check seed organiser shifted labels; one visitor later asked; no recorded resolution. | Sports count difference not theft; “the difference” refers to 8 versus 7. | Theatre note records room move and two questions; confidence not observed. | New practice case, not the held-out studio texts. |
Day 15 fresh check · worked interpretations
Item 1. Correction nominalises or labels an action. The log says Rina swapped the labels on two diagrams at 11:10. It does not say the labels were originally wrong or that the swap corrected them. The word correction already evaluates the action as right; “label swap” is safer. An exact citation is the 11:10 sentence.
Item 2. This change most likely refers to Rina's label swap, but can be made more precise as “the label swap”. “Removed confusion” and “secured understanding” are outcomes the log does not test. One observer's question at 11:20 is recorded; no later test/feedback is supplied. Therefore would assert a result from missing evidence, so it does not repair the claim.
Item 3 exemplar. “At 11:10, Rina swapped the labels on two paper diagrams. Later, one observer asked which diagram matched the switch. The log does not show whether the label swap helped anyone understand the model.” Equivalent two- or three-sentence versions are valid. A useful summary noun such as “the label swap” works after actor/action is named. Do not require the exemplar's exact words.
Next moves: Wrong antecedent → trace a phrase back with two candidate anchors. “Correction” accepted as fact → ask what the source would need to prove the old labels wrong. Result connective overclaim → split event from later effect. Coherent reasoning after read-aloud → record analysis as accessible and check independent reading separately. A missing response due to access interruption is N, not zero ability.
Day 20 fresh check · worked interpretations
Item 1. Reordering packages Jo exchanged cards 2 and 3 at 2:15. This success points vaguely to the exchange or claimed improvement and assumes an outcome; no post-exchange trial, visitor interview or total reader count establishes improved clarity or ended uncertainty. “Later” orders the visitor pause after the exchange but does not prove the exchange caused it.
Item 2. J2's clarity claims a report outcome not in the log. In J3, the visitor traces the blank tabletop, Rae lifts/sets down a card, and Rae says “Maybe start again”; a reader may interpret hesitation, but it is a literary inference from narrated action, not a measured psychological fact. A different supported noun (e.g. “uncertainty”) is valid with exact evidence and cautious wording. Do not treat figurative “room seemed to wait” as a literal measurement.
Item 3 exemplar. “At 2:15, Jo exchanged instruction cards 2 and 3. Later, one visitor paused at card 3 and asked what came next; that question is recorded, but its cause is not. The log does not show whether the exchange improved clarity. In the separate fictional scene, Rae's ‘Maybe start again’ may suggest hesitation.” Accept other clear, limited versions with an actor, anchored reference/summary noun, warranted connective and genre distinction.
Next moves: Actor omitted → unpack reordering; this success vague → replace with a precise noun and ask for evidence; chronology read as causation → contrast “later” with “because”; fiction inferred as clinical fact → add “may suggest” and quote the action; editing loses audience → ask an absent reader what happened first and what is still unknown. A polished style without source fidelity needs source retracing, not praise as accurate reporting.
Observation ledger: record one source phrase, the learner's explanation, I/A/P/N, and one next teaching move for reference chain, connective, abstract-noun function, evidence boundary and revision. These checks have no validated total, classroom pilot or outcome claim. Original SubjectNest staff key © NeuroForgeIO Pty Ltd 2026, CC BY 4.0.