# Prepared teaching examples

These examples were written for the workshop. They are not captured outputs from a live ChatGPT run. Compare your actual response with the original sources, not just these examples.

CSV data row numbers below exclude the header.

## Organization
A17: manifest.csv rows 1-4 lists record, statement, transcript, and one reference. Application record: applications.csv row 1. Staff verifies actual documents.
A18: manifest.csv rows 5-7 lists record, statement, and transcript. No reference listing. Application record: applications.csv row 2. Staff checks intake.
A19: manifest.csv rows 8-12 lists record, statement, two references, and an exception-request file. No transcript listing. Application record: applications.csv row 3. Staff checks intake.
All three IDs and all twelve manifest rows must remain in the inventory. Do not silently discard the exception-request file.

## Administrative review
Rules: program-rules.md sections Required items, Deadline, Exceptions, Decision responsibility.
A17: required types listed (manifest rows 1-4). Recorded 2026-09-28 (applications row 1), within the fictional deadline. Actual receipt and contents remain unverified.
A18: no reference listed (manifest rows 5-7). Recorded 2026-09-30 (applications row 2). Staff checks intake and applicable exceptions.
A19: no transcript listed (manifest rows 8-12). Recorded 2026-10-02 with an exception request (applications row 3). Staff reviews the exception and records. No automatic rejection or eligibility decision.

## Rubric evidence
A17 quote: “I organized 12 weekend tutoring sessions for sixth graders and recruited two volunteers.” Source: statements.md, A17, paragraph 1.
A18 quote: “I set up 12 weekend tutoring days for sixth graders. I got two volunteers to help.” Source: statements.md, A18, paragraph 1.
Both provisionally match anchor 2 of 3 in rubric.md: specific contribution. Neither supplied excerpt provides a supported result for anchor 3. These are applicant-reported claims, not independent verification.
Neutral clarification for both: What result of the tutoring activity can you describe and support?
Human check: verify the exact quote, location, anchor definitions, and equivalent treatment before deciding a final score. Do not rank applicants or infer personal traits.

## Deliberately flawed summary
“Founded a program serving 120 students. Improved grades by 20%. Strong leadership. Score 3 of 3.”
This is an intentionally invented bad example, not a live model output. None of the added claims appears in A17. Correct it to the quoted contribution, provisional anchor 2, and missing outcome evidence. Record the correction.
