Skip to contents

Calculates the coefficient matrices needed for the optimization problem in the combined stratified rank sum test.

Usage

comb_matrix_block(
  Z,
  Y,
  block,
  c,
  methods.list.all,
  scores.list.all = NULL,
  block.sum = NULL
)

Arguments

Z

An n-dimensional binary treatment assignment vector.

Y

An n-dimensional observed outcome vector.

block

An n-dimensional vector specifying block membership.

c

Threshold for the null hypothesis.

methods.list.all

A list of lists of method specifications.

scores.list.all

Optional pre-computed scores.

block.sum

Optional pre-computed block summary.

Value

A list with H elements, each containing a list of B matrices. Each matrix is 2 x (mb+1) containing the number of units and minimum statistic values for different coverage scenarios.