Step 2 of the replication code: cleans the raw baseline survey files and saves the intermediate baseline datasets (demographics, household, household composition, child, mother, father, knowledge, practices, and observations with the decision-making, grit and Rosenberg scales).