R code for “Permutation inference with a finite number of heterogeneous clusters”

This page provides R code for “Permutation inference with a finite number of heterogeneous clusters” [PDF].

Include this file in R with include('hagemann_ap.R'). Alternatively, load it into R directly with

include('https://hgmn.github.io/assets/hagemann_ap.R')

The code supplies two functions:

ApAlphaBar returns α̅ as in Table 1 of the paper for combinations of q₁ and q₀ up to 32 by accessing this list. It has the following arguments:

q1 is the number of treated clusters.
q0 is the number of control clusters.
alpha is the desired level of the test. Default is alpha=.05.
local=TRUE looks for critial value tables in the working directory if set to TRUE. Default is FALSE.

ApTest calculates the test decision as in Algorithm 3.3. It has the following arguments:

x1 is a vector of cluster-level estimates from control clusters.
x0 is a vector of cluster-level estimates from control clusters.
alpha is the desired level of the test.
alphabar will be calculated automatically if alpha is supplied as argument. If alphabar is supplied, then alpha will be ignored.
random=TRUE computes 10⁴ random permutations if the set of all permutations is larger than 10⁴. Default is TRUE.
verbose=TRUE provides a verbal summary of the test decision. Otherwise the value is a 1/0 decision. Default is TRUE.

The code also contains two auxiliary functions for generating permutations.