# Day 44 · Say what the data really tells us

**Try:** “What can we say? What stays unknown?”

**Adult teaching purpose:** Interpret available categorical data and its limits; AC9M2ST01 and AS categorical-data component.

**1 · Prepare and invite.** Have MODEL1 full source/list/table or read it aloud. Prepare two original statements: “maze has most recorded game choices” and “everyone chose maze”. No real child preference, class mandate or population prediction follows from this fictional source. An actual survey needs its own scope, date, willing-response and adult privacy record.

**2 · Model a different small example.** Using unrelated pet-picture tokens (bird3/fish1/no reply2), model “bird has more recorded picture choices than fish” and reject “every family has a bird”. Distinguish a preference answer from actual ownership, and a model from a real survey. An unanswered invitation gives no preference. Explain a tie with another two/two example; neither category has more.

**3 · Guide one attempt.** Ask which MODEL1 statement is supported. Read each frequency and the missing-response status. Maze3/matching2/story1 supports most recorded game choices; it does not support everyone or a claim about other children. Compare maze and matching: one more. Let the child revise a too-large claim in their own words or direct a change.

**4 · Your own useful turn.** Choose MODEL2 or an own reviewed table. Make one useful true statement and one question the table cannot answer. The statement can be spoken/signed/dictated or attached to a private table. A child may choose a tentative game recommendation while acknowledging missing replies; a recommendation is not an instruction that others must join.

**5 · Observe and record.** Keep the actual choice, representation, method and meaning before a correction. Distinguish a source fact, child decision, adult reading/scribing/handling and mathematical hint. Observe whether claims stay with the source, variable and available replies. If they say biggest number means best game, teach data versus value judgment with unrelated shapes. If missing replies are assigned to a category, model uncertainty using a new four-invitation example. If secure, ask how another question or later willing replies could change the decision. Do not fabricate responses to improve a result.

**6 · Teach next and revisit.** Choose just the applicable next teaching in step5, on a different ordinary source; then offer a new question after an actual interval. Record what happened rather than completing all branches. Read the three [participation choices](../ROUTES.md#day-44); [parent offer](../family/home-44.md) is optional. A short useful attempt can replace a long sheet. Full source/rights and curriculum evidence stay in the adult [mapping](../curriculum/MAPPING.md) and [notice](../legal/NOTICE.md).

Original SubjectNest educational material ©2026 SubjectNest contributors. CC BY4.0; full portable conditions and outside exceptions: [notice](../legal/NOTICE.md).