Keep prompt / exact learner words / source seen or heard / adult or peer help / mode / next move. Whole-class answers guide teaching but do not prove independent understanding. Rotate focus learners, then use the fresh Days 5 and 10 checks with each learner. Mark secure / developing / not yet observed per criterion, not a whole-year grade. Treat comprehension, written language, independent reading and speaking mode as different constructs.
| Day | Evidence to catch | Keyed response / next move |
|---|---|---|
| 1 | Headline meaning at two repairs | A=$32, B=$20. A's lower marginal price does not make its total lower here. If confused, compare fixed and changing cost in two colours. |
| 2 | Substantiation at eight repairs | A=$74, B=$80, so A lower by $6 in the model. If a claim has no figures, ask “Which line of the source can a reader check?” |
| 3 | Hybrid structure | Slogan attracts; small print gives terms/formulas. If learner calls one part “fake” solely because it persuades, ask what each actually says. |
| 4 | Actor in shelter proposal | The fictional post proposes removal; it is not a school decision. If nominalisation hides an actor, restore one only if the source supports it. |
| 5 | Independent model note | See Check A. |
| 6 | Reach of 80% | It describes 16/20 at racks, not all students. If learner says the division is wrong, verify arithmetic, then revisit selection. |
| 7 | Full quotation | “beside the bicycle racks” must stay if quoting selection method. If lost, compare shortened/full quote side by side. |
| 8 | Complex sentence | “Although 16 of 20 by the racks reported cycling, that site may overrepresent cyclists.” Main clause stands and condition is honest. |
| 9 | Respectful challenge | Acknowledge reported fraction, challenge population inference and ask for a fairer method. Do not score accent or extroversion. |
| 10 | Fresh source response | See Check B. |
Day 5 check A
Prompt: “Write a 4–6 sentence note to a fictional neighbour considering five repairs. Explain which model plan costs less, show how you know, give one limitation and one question you would ask before a real choice. Text A remains available as the source.”
Key: A(5)=18+7×5=$53; B(5)=10×5=$50; B is lower by $3 in this model. A strong note identifies the flyer as source, distinguishes total from per-repair charge, and says a real decision could depend on parts, quality, eligibility or other omitted conditions. It must not call B universally best. “I would check whether parts are included” is valid. A different opinion about real value can be accepted if no missing fact is fabricated.
Possible answer, not a script to copy: “For five repairs, Plan B has the lower model total. Text A gives A as $18 plus $7 each, so five repairs cost $53; B costs $50. This $3 difference leaves out parts and other terms. I would ask whether both plans cover the same repairs before choosing.”
Day 10 check B
Fresh fictional card: “An imaginary arts-club flyer says, ‘Our survey proves 90% of all students want a new music-room session.’ It quotes ‘18 of 20 wanted it’ but does not say that those 20 were asked at a music-club meeting. In a constructed model of 100 students, 60 are interested; two separate constructed simple random samples of 20 found 11 and 14 interested. No room capacity, accessibility or staffing information is supplied.”
Ask learners to write a note to a fictional school decision group and deliver its main point in 30–60 seconds by an agreed mode. They should: (1) quote the sampling location without losing it; (2) evaluate the 90% claim; (3) use one comparison figure; (4) suggest a better next source or decision step; (5) keep any recommendation bounded.
Answer key: 18/20=90% for meeting attendees, but the meeting selects people likely interested; it does not establish 90% of all students. The known fictional model population is 60/100=60%, and the constructed random samples report 55% and 70%, showing sample variation. The flyer also omits room capacity/access and staff conditions. A sound recommendation could be to gather a fairer sample and check provision before deciding. Do not require one preferred policy choice. A student who notes the model population is known only in this artificial exercise shows strong source awareness.
Four-criterion feedback rubric
| Criterion | Secure in these tasks | Developing | Next teaching move |
|---|---|---|---|
| Claim and source | Names whose claim it is and retains essential source context | Repeats a slogan or clipped quote as fact | Compare headline, full source and missing line; ask who wrote each |
| Substantiation | Exact figure/example supports a specific claim and is labelled as model/sample | Figure is correct but attached to wrong group or unstated source | Use claim–evidence arrows and force a group/source label |
| Reasoning boundary | Separates model result from real choice or sample from population | Says “proves/always/everyone” beyond evidence | Ask what data or condition would change the conclusion |
| Audience and revision | Paragraph/pitch is navigable, respectful and revises from a listener question | Correct pieces are disconnected or a challenge targets a person | Sequence cards; rehearse paraphrase–reason–question privately |
For learners using AAC, dictate/record the learner's chosen words and options offered. If a student heard the source, do not treat their good reasoning as proof of independent reading. If English is an additional language, permit a verified home-language explanation when available and teach task vocabulary separately; no translation is supplied or validated here. If the communication mode blocked a response, change the mode and recheck.
Rights: Original assessment © NeuroForgeIO Pty Ltd 2026, SubjectNest, CC BY 4.0; credit, source and changes required.