block_randomization: Randomized block treatment assignment

In IyarLin/utilityFunctions: A bunch of utility functions for variuos operations

Description

this function peforms block randomization of 2 treatments.

Usage

block_randomization(
  df,
  blocking_vars,
  frac,
  treatment_names = c("control", "treatment")
)

Arguments

df	input data.frame. should include the variables in blocking_vars.
blocking_vars	string vector denoting the names of variables in df to perform block randomization over
frac	a numeric between 0 and 1. the fraction of observations assigned to the 2nd treatment group
treatment_names	names of the treatment groups. default is "control" for the first group and "treatment" for the second.

Value

a vector of treatment assignments

Examples

library(utilityFunctions)
df <- data.frame(var1 = c(rep("A", 99), rep("B", 201), rep("C", 44), rep("D", 56)),
                 var2 = c(rep("E", 25), rep("F", 100), rep("G", 100), rep("H", 100), rep("I", 75)),
                 var3 = rep("J", 400))
blocking_vars <- c("var1", "var2")

df$treatment <- block_randomization(df, blocking_vars, frac = 0.8,
                                    treatment_names = c("treat 1", "treat 2"))

# In the below example, not all blockings can attain an exact fraction of 0.8
tapply(df$treatment, paste0(df$var1, df$var2), function(x) mean(x == "treat 2"))

IyarLin/utilityFunctions documentation built on May 19, 2020, 12:50 p.m.