Replication Materials for “Did Safer Drinking-Water Sources Deliver Safer Water to Drink? Paired Source and Household Evidence from Three National Surveys in Bangladesh, 2012–2025”
Description
This repository contains replication materials for the manuscript “Did Safer Drinking-Water Sources Deliver Safer Water to Drink? Paired Source and Household Evidence from Three National Surveys in Bangladesh, 2012–2025.” The study examines whether drinking-water sources that test microbiologically safe also deliver safe water at the household or point of use, using nationally representative Bangladesh Multiple Indicator Cluster Survey (MICS) water-quality data from 2012–13, 2019, and 2025. The three survey rounds are analysed as independent repeated cross-sections. The harmonized paired analytical sample contains 14,298 households with valid microbiological measurements at both the tested drinking-water source and the household/point of use: 2,302 observations from 2012–13, 6,052 from 2019, and 5,944 from 2025. The primary analytical domain consists of households whose tested source water had no detected E. coli, while the primary outcome identifies E. coli detection in the corresponding household or point-of-use drinking-water sample. The empirical analysis compares source and household water-quality states across survey rounds and examines adjusted prevalence patterns and relevant household, geographic, and socioeconomic heterogeneity. The analysis is observational and repeated cross-sectional; the statistical associations do not by themselves establish a specific causal pathway through which contamination occurred between the source and point of use. The public replication package contains Stata 14-compatible analysis and verification code, variable and model documentation, aggregate statistical tables, figures, captions, validation materials, an input-data schema, and automated result checks. Respondent-level Bangladesh MICS microdata and the harmonized household-level analytical dataset are not redistributed. Researchers wishing to reproduce the complete analysis must obtain authorized access to the relevant Bangladesh MICS survey files through the official MICS data-access process and construct or supply the harmonized analytical input described in the repository documentation. Third-party source materials remain subject to the applicable terms of the original data providers.
Files
Steps to reproduce
Download and extract the replication ZIP while preserving the folder structure and read the separately deposited README.txt. The analysis requires Stata 14.0 or later and does not require user-written Stata packages. Respondent-level Bangladesh MICS data and the harmonized household-level analytical dataset are not redistributed with this repository. Researchers wishing to reproduce the complete analysis must first obtain authorized access to the relevant Bangladesh MICS survey files through the official MICS data-access process. Construct or supply the harmonized analytical file using the variable structure documented in data/analysis/MICS_Water_AnalysisReady_SCHEMA.csv and place the authorized file at data/analysis/MICS_Water_AnalysisReady.dta. Open Stata, set the current working directory to the extracted replication-package root, and run do "RUN_ALL.do". The workflow reproduces the statistical analyses, aggregate manuscript tables and figures and generates output/verification_report.txt. A successful reproduction regenerates the principal manuscript tables and figures and completes the supplied verification checks. The repository reproduces the statistical analysis from the authorized harmonized analytical input; it does not reconstruct that harmonized file directly from the original round-specific MICS microdata.
Institutions
- Jagannath UniversityDhaka Division, Dhaka