HarmonicWtdMean


The level of treatment variable (e.g., "treated" or 1) that indicates treated observations



Blocking/stratification variable (vector)


block


  Additional arguments (included for compatibility but ignored)
  


Calculates the weighted average of the block-specific differences of means, where the weights are proportional to the harmonic mean cluster size in the block (Donner and Klar, 1993; Bowers, Fredrickson and Hansen, 2010). The result is intended to be used as a test statistic in a permutation test.


 ~kwd1 

 ~kwd2 

An implementation of nonparametric combination of hypothesis tests.
This package performs nonparametric combination (Pesarin and Salmaso 2010),
a permutation-based procedure for jointly testing multiple hypotheses. The
tests are conducted under the global "sharp" null hypothesis of no effects,
and the component tests are combined on the metric of their p-values. A
key feature of nonparametric combination is that it accounts for the
dependence among tests under the null hypothesis. In addition to the
"NPC" function, which performs nonparametric combination itself, the
package also contains a number of helper functions, many of which calculate
a test statistic given an input of data.

HarmonicWtdMean: Block-Specific Mean Differences Weighted by Harmonic Mean Sample Size

Description

Usage

Arguments

References