The structure of the cleaning files is modular: one can harmonize cross-sectionally across countries, or longitudinally within one survey, or both longitudinally and cross-nationally.
Download the survey data & organize according to the "Read me" file.
Use the "set file locations" to set the directories of where all the survey data was downloaded & the location of cleaning do files.
Use the "master" do file to select which surveys & topics to harmonize. You must include the topic Full_Sample_IDs. This establishes the universe of children <=5 at some point.
If you do not want to harmonize every survey round, you need to change the rounds it cycles through within each topical cleaning file for the country.
Each do file more or less has the following structure:
A loop through all rounds selecting all variables used to generate the variables of interest and renaming them to be consistent across rounds
A loop through all rounds manipulating the variables into their final form
All rounds are appended together and a long form of the dataset is generated.
The data is reshaped wide so there is one observation per child and another dataset is saved.
Stata do files are available here
Suggested citation: Reynolds, S.A. 2023. Stata Code for Project CHILD Harmonized Data. https://sites.google.com/view/projectchild