BAEssd-package {BAEssd} | R Documentation |

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).

############################################################ # 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)

[Package *BAEssd* version 1.0.1 Index]