# Log

• 2026-10-30 P0 contract written. P1 audit: clean data, no exclusions.
• 2026-10-30 P2 baseline OLS iid SE computed (in analysis.py). Decision: main SE = CR1 cluster-robust by school, t(G−1=39) reference; ICC via one-way ANOVA on OLS residuals. No randomness.
• 2026-10-30 P3 analysis.py run. P4 all checks pass. P5 report and claims written.