Employes a Bayesian average error based approach to sample size determination. Several functions are included for sample size calculation for common designs in clinical trials including one- and two-sample binary and normal responses. These functions are grouped in "suites" that pertain to each type of example.
Package: | BAEssd |
Type: | Package |
Version: | 1.0.1 |
Date: | 2010-06-18 |
Suggests: | mvtnorm |
License: | GPL-2 |
LazyLoad: | yes |
Before any calculations can be made, first identify the example of interest and the corresponding suite of functions:
binom1.1sided | One sample, binary response, one-sided hypothesis. |
binom1.2sided | One sample, binary response, two-sided hypothesis. |
binom2.1sided | Two indepednent samples, binary response, one-sided hypothesis. |
binom2.2sided | Two independent samples, binary response, two-sided hypothesis. |
norm1KV.1sided | One sample, normal response, known variance, one-sided hypothesis. |
norm1KV.2sided | One sample, normal response, known variance, two-sided hypothesis. |
norm1UV.2sided | One sample, normal response, unknown variance, two-sided hypothesis. |
norm2KV.2sided | Two independent samples, normal response, |
known variance, two-sided hypothesis. | |
After selecting the suite of functions of interest, the suite must be
generated with appropriate parameters. Then, the corresponding ssd
function can be used to calculate the sample size. The two primary constraints
for choosing a sample size are the bound on the Total Error to maintain and
the weight given to controlling the Average Type-I Error (as opposed to the
Average Type-II Error).
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 | ############################################################
# Calculate the sample size required for a one-sample
# normal experiment with known variance (sigma2=25) with
# the hypothesis of interest being
# H0: theta==0 vs. H1: theta!=0
#
# where theta is the mean of the normal distribution. For
# details on the prior used, see documentation for
# norm1KV.2sided().
# generate suite of functions
f1 <- norm1KV.2sided(sigma=5,theta0=0,prob=0.5,mu=2,tau=1)
# attach suite
attach(f1)
# calculate sample size for TE bound of 0.25 and weight 0.5
ssd.norm1KV.2sided(alpha=0.25,w=0.5)
# detach suite
detach(f1)
############################################################
# Calculate the sample size required for a two-sample
# experiment with a binary response in which the hypothesis
# of interest is
# H0: p1==p2 vs. H1: p1!=p2
#
# where p1 is the response rate for group 1 and p2 is the
# response rate for group 2, independent samples. For
# details on the prior used, see documentation for
# binom2.2sided().
# generate suite of functions
f2 <- binom2.2sided(prob=0.5,a0=1,b0=1,a1=1,b1=1,a2=1,b2=1)
# attach suite
attach(f2)
# calculate sample size for TE bound of 0.25 and weight 0.5
# - here the log marginal distribution (logm) is part of the suite.
ssd.binom(alpha=0.25,w=0.5,logm=logm,two.sample=TRUE)
# detach suite
detach(f2)
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.