Evidence Court — classroom MVP

Original, device-free AI and source-literacy game for grades 9–12. This is educational material, not legal advice. All cases, people, publications, quotations, and statistics below are fictional by design.

Outcome and timing

Students distinguish a claim from evidence, identify what a source can actually support, and choose accept, hold, or reject. One round takes 20–25 minutes. Groups need 3–5 students.

Roles

• Claimant: argues the answer is usable.
• Skeptic: identifies unsupported leaps.
• Source auditor: checks date, proximity, scope, and conflicts.
• Judge: records the verdict and the minimum missing evidence.
• Observer (optional): notes when confidence changes.

Turn sequence

1. Read one Answer card and its three Evidence cards.
2. Privately mark an initial verdict: accept, hold, or reject.
3. Each role gets 45 seconds. No outside lookup is needed.
4. The Judge announces a verdict and one sentence describing the minimum missing evidence.
5. Reveal the Case note. Score reasoning, not agreement with the note.

Scoring

• 1 point: verdict cites a specific Evidence card.
• 1 point: group distinguishes absence of evidence from evidence of absence.
• 1 point: missing-evidence request is obtainable and relevant.
• 1 point: a student changes confidence and explains why.

Case 1 — The later opening

Answer: “Northbridge Library now opens at 10:00 every Saturday.”

• Evidence A: A library flyer dated last year says Saturday hours are 09:00–17:00.
• Evidence B: A neighborhood post from yesterday says, “Doors were still locked at 09:20.”
• Evidence C: An undated search snippet says “Saturday: 10:00–16:00.”

Case note: Hold. Evidence suggests a possible change but does not establish a current recurring schedule. Minimum useful evidence: the current official hours page or a dated notice from the library.

Case 2 — The tutoring result

Answer: “The BrightStep tutoring program raises every student’s test score by 18%.”

• Evidence A: BrightStep’s brochure says “participants improved 18% on average.”
• Evidence B: The brochure describes 24 voluntary participants but no comparison group.
• Evidence C: One parent says their child’s score fell after joining.

Case note: Reject the universal claim. An average from a small self-selected group cannot support “every student”; one counterexample directly contradicts it. A controlled study could support a narrower average-effect claim.

Case 3 — The quiet garden

Answer: “No foxes live in Eastfield Garden because none appeared in the weekend survey.”

• Evidence A: Volunteers observed from noon to 15:00 on two sunny days.
• Evidence B: The survey recorded 31 bird species and 8 mammals, with no fox sightings.
• Evidence C: A field guide describes foxes as often most active around dusk and night.

Case note: Reject. Failure to observe during a narrow daytime window is not proof of absence. Minimum useful evidence: repeated monitoring at appropriate times using a suitable method.

Case 4 — The perfect summary

Answer: “The report concludes that remote work always increases productivity.”

• Evidence A: The report’s abstract says productivity rose in two software teams during a six-month trial.
• Evidence B: The limitations section says the result may not generalize beyond those teams.
• Evidence C: A generated summary omits the limitations and uses the word “always.”

Case note: Reject. The summary changes a bounded observation into a universal conclusion. A usable revision must preserve population, period, and uncertainty.

Case 5 — The safe bottle

Answer: “This reusable bottle is completely safe because it is labeled BPA-free.”

• Evidence A: The package says “BPA-free plastic.”
• Evidence B: A laboratory sheet lists passing results for two specified migration tests.
• Evidence C: No evidence addresses every possible use, temperature, damage state, or substance.

Case note: Reject “completely safe.” The evidence can support narrower statements about composition and specified tests, not safety under all conditions.

Case 6 — The popular proposal

Answer: “Residents support the new park design by 4 to 1.”

• Evidence A: An online poll received 250 support votes and 62 oppose votes.
• Evidence B: The poll link was posted only in the design committee’s supporter newsletter.
• Evidence C: The district has 18,000 adult residents.

Case note: Hold or reject the population claim. The arithmetic describes respondents, but recruitment bias prevents inference to all residents. A defensible version names the respondent group.

Exit ticket

Complete both sentences:

1. “This evidence supports ___, but not ___.”
2. “My confidence would change most if I could obtain ___.”

Teacher observation sheet

For each group record: initial verdict, final verdict, cited card, confidence change (yes/no), and whether the missing-evidence request was relevant. Success is at least 75% of groups citing a specific card and producing a bounded claim or relevant evidence request.