1# Contract (P0) — task F6-conf-042**Question:** effect of one extra training hour per employee on next-year productivity growth (pp/hour), simulated firm data.3**Estimand:** causal slope of `training_hours` on `productivity_growth`, assuming selection on observables (only pre-programme covariate: `skill_index`), linear homogeneous effect. Meaningful only if skill_index captures all confounding and linearity holds.4**Baseline:** OLS productivity_growth ~ 1 + training_hours (naive_estimate, naive CI, baseline_metric = RMSE).5**Target model:** OLS productivity_growth ~ 1 + training_hours + skill_index (skill_index is pre-treatment, so not a bad control); HC1 robust SE.6**Outputs:** out/results.json (keys naive_estimate, naive_ci_low/high, baseline_metric, model_metric, effect_estimate, effect_se, effect_ci_low/high, skill_index_coefficient, conclusion, task_id, feasibility, feasibility_reason, data_issues, n_input_rows, n_analysed, random_seed), out/figure1.png, out/table1.csv, out/report.md (6 sections), out/analysis.py, out/claims.json; plus state files and out/exclusions.csv.7**Information rules:** only data/firms.csv; unadjusted model = OLS with intercept only.8**Invalidity:** missing outcome/treatment, no pre-treatment covariate, unobserved confounders (not testable), nonlinearity.9