Skip to main content
. 2015 Jun 4;10(6):e0128933. doi: 10.1371/journal.pone.0128933

Fig 1. Flow diagram of pooled DHS dataset illustrating where countries and observations are lost due to missing variables or values in analysis of (a) main cooking fuel, (b) cooking location and (c) stove ventilation.

Fig 1

a Benin, Burkina Faso, Cameroon, Ethiopia, Gabon, Ghana, Guinea, Kenya, Madagascar, Mali, Malawi, Mozambique, Namibia, Niger, Rwanda, Senegal, Tanzania, Uganda, Zambia, and Zimbabwe. b Ethiopia, Ghana, Kenya, Madagascar, Malawi, Namibia, Uganda, Zambia, and Zimbabwe. c Ghana, Madagascar, Malawi, Namibia, Uganda, and Zambia. d Missing values in explanatory variables: Maternal smoking, crowding, time to nearest water source, shelter index, wealth index, vaccination index, paternal education, maternal education, possession of health card, child sex, child age, birth order, breastfeeding duration, stunting, geographic location, sex of household head, age of household head. e Variables not selected in final model: Crowding, wealth index, possession of health card, breastfeeding duration, sex of household head, age of household head. f Three countries could be added because crowding was not selected in the final model and it was the only missing variable in Burkina Faso, Guinea, and Senegal. g Missing values in explanatory variables: Maternal smoking, time to nearest water source, shelter index, paternal education, maternal education, child sex, child age, birth order, stunting, vaccination index, geographic location.