Skip to content

[R] package for the computation of optimal sample sizes in the sense that these maximize the expected reward of a two-stage design for a finite population.

Notifications You must be signed in to change notification settings

Nth-iteration-labs/ssev

Repository files navigation

ssev

The goal of ssev is to compute optimal sample sizes for two-stage RCT designs when incorporating the size of the population. The basic idea is simple: given an estimate of the effect size of the difference between two groups, and an estimate of the population size N, we can compute the optimal size of an RCT that maximizes the expected outcome over both the RCT and the resulting guideline that is used for the remainder of the population. The package computes the associated sample size for a given significance level when comparing two means (with equal or unequal variances assumed) or when comparing two proportions.

Example

The following code computes the sample size required for a population of 10000 units when assuming equal variances and a Cohen's d of .8:

compute_sample_size(means=c(0,.8), sds=1, N=10^5)

About

[R] package for the computation of optimal sample sizes in the sense that these maximize the expected reward of a two-stage design for a finite population.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages