Fig. 1From: The impact of nonresponse in different survey stages on the precision of prevalence estimates for multi-stage survey studiesFlowchart of the simulation and data analysis procedure. This flowchart illustrates an example of the process to create 1000 datasets with 5% nonresponding schools, classes, and students, respectively, and the process to calculate and compare the mean margin of errors for the data with nonresponding schools, classes, and students, using current smokers as the survey outcome indicator. The simulation and data analysis were also conducted using other non-response rates (i.e., 10%, 20%, 30%, 40%, 50%, and 60%) and other survey question indicatorsBack to article page