Randomization-Based Inference

Randomization-Based Inference for customized experiments. Computes Fisher-Exact P-Values alongside null randomization distributions. Retrieves counternull sets and generates counternull distributions. Computes Fisher Intervals and Fisher-Adjusted P-Values. Package includes visualization of randomization distributions and Fisher Intervals. Users can input custom test statistics and their own methods for randomization. Rosenthal and Rubin (1994) .

Counternull

Counternull package allows users to conduct Randomization-Based Inference for customized experiments.Users may use the package to compute Fisher-Exact P-Values alongside null randomization distributions.Additionally, users can retrieve counternull sets, generate counternull distributions, compute Fisher Intervals, and Fisher-Adjusted P-Values. The package may be used on data of any size and distribution including usage with custom made test statistics.

Installation

You can install the released version of Counternull from CRAN with:

install.packages("Counternull")

Usage

Examples of functions that can be used in Counternull Package:

library(Counternull)
y = sample_data$turn_angle
w = sample_data$w
n_r = create_null_rand(y, w, sample_matrix, test_stat = c("t"))
summary(n_r)
#> Observed test statistic: 1.88171 
#> Number of extreme test statistics: 56 
#> P-value: 0.056 
#> Alternative: two-sided
plot(n_r)

n_r = create_null_rand(sample_data$turn_angle, sample_data$w,
sample_matrix, test_stat = c("diffmeans"))
c = find_counternull_values(n_r)
summary(c)
#> Counternull Set (Positive): [ 5.782512 , 5.817145 ] 
#> Counternull Set (Negative): [ -5.851778 , -5.841883 ]
plot(c)

License

MIT