# Data audit

• data/schools.csv SHA-256 0f1d8378fff2f0898762feaa08dd94d9ad04156dad6330572d98997c9005c88a (matches bundle.json)
• 1000 rows × 4 cols: school_id int (100–139), student_id int (1–1000, unique), treated int {0,1}, score float (points; 35.72–64.17, mean 50.80, sd 4.48)
• Missing values: none. Duplicate rows/student_ids: none.
• 40 schools × 25 students (balanced); treated constant within every school; 20 treated / 20 control schools.
• No anomalies; no availability-rule violations (single permitted input). No rows excluded (see out/exclusions.csv).