# Learner studio · Paper Screen inquiry

**Our evidence is fictional.** Read [Sources A–E](SOURCE-PACKET.md). The numbers are invented sensor units, not results from a real room or a safety test. Use the [print and text aids](print/TEXT-ALTERNATIVES.md) if helpful. Show which source supports each statement. You can write, speak, sign, use AAC, draw a high-contrast diagram, or direct a scribe; keep the same reasoning.

## Week 3

### Day 11 · Read the whole poster
Circle the headline and the quieter sentence in Source A. Who wrote it, to whom, and why? Compare its four-layer **prediction** to Source B's four-layer **record**. Write a two-sentence replacement headline and one unanswered question. Choose an annotated page, claim/record cards, or a spoken/signed source comparison. Try the same routine on an entirely made-up stage-light advert at home.

### Day 12 · Write the conditional model
On the [model table](print/model-and-data.pdf), use `P(n)=160×2^-n` to predict n=0 through 4. Explain `2^0` and `2^-2` with marks, factors or words. Put the recorded B values in a different column. What would the model say at an **untested** n=5? Mark it prediction only. Choose paper strips, raised cards, or symbolic working. At home explain `2^-3` with eight equal marks.

### Day 13 · Name the difference
Calculate `record − prediction` for each B row. Plot paired values or arrange number tiles. What pattern can you see, and what can't you know from B? Write one claim that B can support and one question for a better test. Choose tactile dots, number tiles, or a written table. At home invent a tiny prediction/record pair and label its difference.

### Day 14 · Plan a repeat someone could follow
Ask how a sensor value varies with number of layers **for this setup**. Specify what changes, what is read, what stays fixed, how many readings, how to handle background light and how to record results. Use the [method mat](print/retest-method.pdf). Peer-review a partner's steps, then repair an ambiguous step. Choose procedural cards, control/measure tokens with dictation, or a written protocol. Write `not observed` wherever your class has not made an approved real observation. No equipment is needed at home; tell someone how a fair test would work.

### Day 15 · Fresh transfer check
Your teacher will provide **Check A** separately. Make a first independent response with your available access tools. Afterwards, reflect on how you distinguished assumption, calculation, source and limit. At home, explain why a predicted value is not a measured result.

## Week 4

### Day 16 · Read repeats carefully
In D compare n=0 and n=4 across the three invented rounds. What values are close? Does that fix the mismatch with the half-model? Give one within-record conclusion and one missing method detail. Choose number strips, a raised dot plot, or a written/signed comparison. At home invent three close numbers and one changed condition you would test.

### Day 17 · Try another factor
Explore `P(n)=160×r^n` with `r=.50` and `.55`, using [the offline tool](code/panel_model.py) or a paper calculator table. Compare the average **absolute** gap from B. A smaller gap fits these invented rows better; say why it doesn't prove that factor is a property of all paper. Use a keyboard, teacher-read output with repeated-factor cards, or a paper table. At home draw two repeated-factor steps with a factor of your choice.

### Day 18 · Find the actor hidden by a noun
Read C. Who or what is missing behind “the reduction”? Trace “dependable” to a source or mark it unsupported. Rewrite C for a fictional manager, then as a concise attributed table caption. You may keep an abstract noun if its meaning is clear. Choose annotation, actor/action/result/limit cards, or dictated editing. At home rewrite a made-up sentence that claims “the improvement succeeded” without evidence.

### Day 19 · Make a responsible two-panel brief
Source E asks about a **small supervised retest**, not approval for public use. In [the canvas](print/recommendation-canvas.pdf), make one panel with a numerical comparison and another with your decision, limitation and precise next check. Choose hand-drawn panels, tactile layout with exact-word scribe, or typed panels plus a spoken/signed explanation. Ask a peer what the first glance seems to claim; revise any overstatement. At home write a no-cost `because / however / next` note about fictional data.

### Day 20 · Fresh reproducibility check
Your teacher will provide **Check B** separately. Save your first independent response before discussion. Then mark one fact, one model prediction and one unknown. At home explain the difference between repeated fictional values and a reproducible real investigation.

## When you finish

Read your brief as the fictional manager: can you find the model assumption, exact record, source, missing method detail and safe next step without guessing? If not, revise. You may change format without changing the evidence.