# Data audit (P1)

File: data/design.csv — SHA-256 2f459c539378cbe54d830c18af466f2fa4230ce48a609aa6464f8334a0d36ecb (matches bundle.json)

• 7 data rows x 3 columns (parameter, value, description); header present. No missing values, no duplicate parameters.
• Parameters: n=30 (series length y_1..y_n), rho=0.8, sigma=1.0, intercept=yes, initial_condition=stationary (y_1 ~ N(0, sigma^2/(1-rho^2))), n_replications=20000, seed=20261029.
• Units: dimensionless (sigma in units of y). No dates (simulation design only).
• Anomalies: none. |rho|<1 so stationary start is well defined. Value column mixes numeric and string — parsed per-parameter.
• Availability: only input; nothing violates the rules. Note: regression uses n-1=29 effective observations; formula evaluated at n=30 per task wording.
• No filtering of input rows; the only potential exclusion would be non-finite rho_hat replications (ledger in out/exclusions.csv, computed by analysis.py).