quick and dirty code for systematic censoring

censor.R

library(tidyverse)

# define the censoring function 
censor_smoking <- function(df) {
  
  # (this code is awful I am so sorry)
  df <- df %>% 
    rowwise() %>% # ugh... rowwise
    mutate(keep_case = case_when(
      (sex.factor == "Male" & runif(1) < .3)   ~ 0, # censor male with prob .3
      (sex.factor == "Female" & runif(1) < .6) ~ 0, # censor female with prob .6
      TRUE ~ 1 # otherwise don't censor
    )) %>% ungroup()
  
  # do the censoring
  df$smoking[df_miss$keep_case == 0] <- NA # censor
  df$keep_case <- NULL
  
  # convert to pure df
  df <- as.data.frame(df)
  return(df)
  
}


# small data set to use as an example
df <- tribble(
  ~id,	~sex.factor,	~smoking,
  1, "Male", "Non-smoker",
  2, "Male", "Non-smoker",
  3, "Female", "Non-smoker",
  4, "Female","Smoker",
  5, "Male", "Non-smoker",
  6, "Female", "Smoker",
  7, "Male", "Smoker",
  8, "Male", "Non-smoker",
  9, "Male", "Non-smoker",
  1, "Female", "Non-smoker",
  1, "Female", "Non-smoker",
  1, "Male", "Non-smoker",
  1, "Male", "Non-smoker",
  1, "Male", "Non-smoker",
  1, "Male", "Smoker",
  1, "Male", "Non-smoker",
  1, "Male", "Non-smoker",
  1, "Male", "Smoker"
)

# now actually use it to overwrite the original
df <- censor_smoking(df)
print(df)