Estimates power accounting for realistic missing data patterns.
Usage
run_power_sim_with_missingness(
distribution,
params,
n_per_group,
effect_size,
na_rate = 0,
mnar_score = 0,
alpha = 0.05,
test = "wilcoxon",
n_sim = 1000
)Arguments
- distribution
Distribution name
- params
List of distribution parameters
- n_per_group
Number of samples per group
- effect_size
Fold change for treatment
- na_rate
Proportion of values that will be NA
- mnar_score
MNAR intensity (0 = MCAR)
- alpha
Significance level
- test
Statistical test to use
- n_sim
Number of simulations