bpbounds-package {bpbounds}R Documentation

bpbounds: Nonparametric bounds for the Average Causal Effect due to Balke and Pearl and extensions


This package implements the nonparametric bounds for the average causal effect defined by Balke and Pearl, Bounds on Treatment Effects from Studies with Imperfect Compliance, JASA, 1997; and some extensions.


The functions implement bounds for the situation where each of the outcome, treatment/phenotype, and instrumental variable are binary; and additionally for when the instrument has 3 categories (e.g. a single genotype under an additive model in a Mendelian randomization study).

The package implements bounds for when the three variables are measured in the same study (trivariate data) and when the outcome and instrument are measured in one study and the treatment/phenotype and instrument in another sample (bivariate/two sample data).

[Package bpbounds version 0.1.5 Index]