1# Data audit (P1)2File: data/design.csv — SHA-256 b1a17a529973e7994bbedad5ec1bc73fe968c777301ca8745657ea300e3c74f0 (matches bundle.json).3- 30 rows × 3 columns: i (int64, 1..30, unique, consecutive), x (float64, dimensionless, range −1.3942..1.9919, mean 0.2823, sd 0.8212), sigma (float64, dimensionless, range 0.6034..1.7903, all > 0).4- Missing values: 0. Duplicate rows: 0. Duplicate x: 0. No dates (synthetic design).5- Notable feature (not an error): sigma is strongly increasing in |x| (corr(x², sigma²) = 0.987), i.e. heteroskedasticity loaded on high-leverage points — the setting where conventional SE is expected to be biased downward.6- Availability: the only input; it is the design itself, no rule violations. No exclusions applied (exclusions.csv records a no-op step).7