This is a public formative key, not a secure exam key. It is teacher-targeted but anyone with the URL can read it. Collect and record the first child-controlled response before feedback. If earlier access to the case matters, change the source pictures and values, calculate a new key, and record that local variation. Do not score speed, neatness, speech, handwriting, hearing, motor action or a forced statement that the child's mind changed. Note any adult mathematical hint separately from neutral read-aloud, AAC or child-directed movement.
Check A · Day 185 · Worked answer
Teacher source from the check sheet: 13 invented craft labels, 9 ARC and 4 BAR. There must be one source-to-row placement or mark per picture. A complete sort gives 9 ARC, 4 BAR, and 13 total. Pair four ARC/BAR pictures; 5 ARC pictures remain unmatched, so ARC is more by 5 in this supplied sheet. A counted comparison 9 > 4 is equally valid. The source boundary is these thirteen drawn labels; “most craft shops use arcs” is unsupported.
| Evidence seen | Interpret and act next |
|---|---|
| Thirteen distinct source pictures accounted for, both category counts 9/4, ARC more with count or pairing, and source-limited statement | Target demonstrated for this one set. Next give an unfamiliar category pair with a changed layout, and check again later. |
| Right comparison but a skipped or twice-counted picture | Treat as an unstable one-to-one trail. Use a DONE column and fresh 5/3 source; child moves each picture once. Do not replace their reasoning with an adult tally. |
| Sort is complete but says BAR is more because its row is physically longer | Pair pictures or compress/spread rows while preserving the source; ask what actually changed. |
| Makes a claim about real shops/people | Ask “Which exact pictures did we count?” The source sentence is a separate evidence target. |
| Neutral access support used | Count mathematical decisions made by the child. Record any adult category/count cue and repeat with a new source. |
Check B · Day 190 · Worked answer
Story 1 source: Row A is 12 compact identical marks; row B is 8 more widely spaced identical marks. One-to-one pairing leaves 4 unmatched A marks. The checked comparison is A has more by 4, even though B occupies more paper width. If a child predicted A initially, the recheck confirmed it. If they predicted B, the recheck supports a revision. Do not infer a mind change from the answer alone.
Story 2 source: A fictional display starts with 4 broad icon cards and 5 more are added; a separate claim card reads 8 at the end. The visible model gives 4 + 5 = 9. Check by counting all nine exactly once or by matching START four, ADD five and END nine. The display card needs 9, not 8. This is a model of an invented display, not an observed stock count.
| Evidence seen | Interpret and act next |
|---|---|
| Rows counted/paired as 12/8 with A more and a reason; visible 4 plus 5 model gives 9 and corrects 8; child identifies a confirmed or revised first idea honestly | Two sampled targets demonstrated. Transfer later to fresh layouts and a take-away story; no broad mastery claim. |
| Picks spread B as more even after count | Pair one to one, then redraw equal-width cells around the same marks. Ask which act would change the number: add/remove a mark. |
| Says 9 but cannot show a start and added part | Hide the claim and build a new add-to case (for example 2 + 3) with separate areas before combining. |
| Counts START twice or loses an added card | Place START and ADD in labelled frames; one moved card gets one final count mark. Reset the source if any physical piece is lost. |
| Says “I changed my mind” because asked, despite a correct first prediction | Validate “I checked and confirmed it.” Authentic revision is evidence-dependent, not a performance requirement. |
Teacher decision rule: Use the action and source trail, not only the final numeral, to plan the next case. Count/math targets are partial: an accurate response on these two public paper cases does not establish the whole ACARA description, classroom effectiveness, local syllabus approval or individual access suitability. Store real child work privately; no real student data belongs in this repository.