# Data audit (P1)

Schema: i (int64, 1..30, unique, sequential), x (float64, dimensionless, range -1.3942..1.9919, mean 0.2823, sd 0.8212), sigma (float64, dimensionless, range 0.6034..1.7903, all > 0). No dates (synthetic fixed design). Missing values: 0. Duplicate rows: 0. Anomalies: none. Note: sigma increases with |x - c| (roughly quadratic in x; largest at extreme x), i.e. heteroskedasticity is concentrated at high-leverage points — the setting where conventional SEs understate variability. Availability: the file is the only permitted input; no violation. No exclusions applied.