# Contract — F6-dev-01

Question. Does the cross-country Preston curve (life expectancy vs log real income per capita) hold in the most recent usable year of WDI data, and how strong is it?

Estimand. β in the linear projection LE_i = α + β ln(GDPpc_i) + ε_i over countries/economies (equal weights) in the analysis year. LE = life expectancy at birth (years); GDPpc = GDP pc PPP const. 2021 intl $. Descriptive/associational only. Meaningful as a best linear predictor; robust (HC) SE for inference treating countries as draws from a superpopulation (questionable: countries are close to a census).

Rules. Exclude entities with region == "Aggregates" in wb_entities.csv. Analysis year = most recent year in 2015–2023 with both indicators observed for >=150 countries. Missing-indicator countries in that year excluded and counted (n_excluded_missing); no imputation. Natural log. Only the provided files.

Outputs. out/results.json (analysis_year, n_excluded_missing, baseline_metric, model_metric, pearson_r, r_squared, effect_estimate, intercept, effect_se, effect_ci_low, effect_ci_high, conclusion, task_id, feasibility, feasibility_reason, data_issues, n_input_rows, n_analysed, random_seed); out/figure1.png; out/table1.csv (income_level, n_countries, mean_log_gdp_pc, mean_life_expectancy); out/report.md (6 sections); out/analysis.py; out/claims.json; plus protocol files (state/, exclusions.csv).

Invalidity / impossibility. No year with >=150 complete countries; entities in indicators not matchable to metadata (could leak aggregates); non-positive GDP (log undefined); duplicates of entity-year.