get_arsenic_data = function(){
    fn = '../../Data/arsenic.txt'
    DF = read.table(fn, header=FALSE, skip=0, col.names=c('Age', 'Sex', 'Drinking_Use', 'Cooking_Use', 'Arsenic_in_water_ppm', 'Arsenic_in_nail_ppm'), fill=TRUE)
    DF$Sex = factor(DF$Sex, levels=c(1, 2), labels=c('Male', 'Female')) ## encode the factor variables 
    DF$Drinking_Use = factor(DF$Drinking_Use, levels=1:5, ordered=TRUE)
    DF$Cooking_Use = factor(DF$Cooking_Use, levels=1:5, ordered=TRUE)
    DF[is.na(DF)] = 0
    DF
}