drcmbote/drcmbote-paper2-ownp-did-ethiopia-

# Secular Trends in Child Stunting and Policy Attribution Under Ethiopia's One WASH National Programme Reproducibility package for the analysis reported in: > Secular Trends in Child Stunting and Policy Attribution Under Ethiopia's One WASH National Programme: A Six-Round Generalised Intensity Difference-in-Differences Analysis, 2000-2024. Submitted to *The Lancet Global Health*. This repository contains the R and Python scripts used to construct the pooled six-round analytic panel and to produce every table and figure in the manuscript and its Supplementary Appendix. It does not contain raw data (see **Data access** below). This is Paper 2 of a two-paper series; Paper 1 (zone-level water point functionality and child nutritional outcomes, 2024-25 EDHS) has its own reproducibility repository at drcmbote/paper1-wash-functionality-ethiopia. ## Data access - **Ethiopian Demographic and Health Survey (EDHS)** individual child, woman, and household recode files for all six rounds used (2000, 2005, 2011, 2016, 2019, 2024-25) are available on registration at dhsprogram.com, subject to the DHS Program's standard data use agreement. Redistribution of the raw DHS extracts is not permitted under that agreement, which is why they are not included here. Analysts who have obtained all six rounds under the DHS data use agreement can reproduce the pooled analytic panel (`panel_5rounds_child.csv`, referenced by the scripts below; retains its original five-round filename for continuity even though it was extended to six rounds) using `build_panel_5rounds.py`, which documents every round-specific data quirk handled (wealth-index and anthropometry merges for 2000/2005, `b19` reconstruction, region-code harmonisation across rounds, JMP improved/unimproved keyword classification, SNNPR post-2020 subdivision reunification, and the 2024-25 Gambella spelling fix). ## Script-to-output mapping | Script | Produces | |---|---| | `build_panel_5rounds.py` | Pooled six-round analytic panel …