In stefvanbuuren/RRR: Replacement, Reduction and Refinement

Summary

The objective to make two comparisons of interest:

1. Compare the mean of the induced group to the mean of the control group;
2. Compare the mean of the induced group to the means of r ntreated treatment groups.

The sample size calculation is made for the one-sided two-group t-test with equal variances and unequal sample sizes. It is assumed that r ntreated + 1 tests are performed. The level of significance is set to r alpha. The power in each test should be equal to or higher than r power1.

The calculation uses the following means and standard deviations:

musdval[1:3,]

where the mean and standard deviation for the treatment condition correspond to a reduction of r musdval[4]%.

The optimal symmetric experimental design, with groups of equal size, requires r ifelse(table.sym[1,1]=='-', paste('more than', (ntreated+2)*30),table.sym[1,1]) experimental subjects. The smallest asymmetric experiment requires `r ifelse(table.asym[1,1]=='-', paste('more than', (ntreated+1)*30+2),table.asym[1,1])`` experimental subjects.

The five most efficient designs are

print(table.asym[1:5,], row.names = FALSE)

r date() TNO 3R Sample Size Calculator

---

stefvanbuuren/RRR documentation built on Sept. 8, 2020, 12:11 a.m.