Deconstructing the (un)affordability of clean cooking fuels: evidence from a randomized trial in rural Tanzania

Participant level data for a yearlong stepped-wedge RCT that provided the intervention of a lockbox. After an initial baseline survey, followup surveys occurred at household visits every two weeks (visits 1 to 26). Each participant was randomly assigned to a time point to roll over into treatment, and thus had different total time in control/treatment periods. The df_all excel has outcome data in longformat. The df_baseline excel has the baseline covariates, and df_refill has each participant's total refills for each individual's time in control and treatment (two observations per participants).