bayes_min_n: Calculate minimum sample size for one-stage Bayesian test

In njjms/nicks: Basic Functions for Biostatistics

Calculate minimum sample size for one-stage Bayesian test

Description

This function searches over the range specified by min_n to max_n for the smallest sample size that achieves a desired power (e.g. .8). If desired power is not reached, function defaults to returning the maximum allowable sample size.

Usage

bayes_min_n(
  req_p = 0.995,
  true_p = 0.995,
  gamma = 0.8,
  power = 0.8,
  alpha = 0.05,
  prior_alpha = 0.5,
  prior_beta = 0.5,
  min_n = 500,
  max_n = 1800,
  step_size = 10
)

Arguments

req_p	observed count
true_p	total count
gamma	posterior probability threshold needed to reject, defaults to .8
power	Desired power, defaults to .8
prior_alpha	alpha parameter of beta prior, defaults to .5 (Jeffrey's prior)
prior_beta	beta parameter of beta prior, defaults to .5 (Jeffrey's prior)
min_n	smallest n to search
max_n	maximum possible n to search
n1	first stage sample size, defaults to 200
n2	second stage sample size, defaults to 800

Value

list containing minimum selected sample size and associated probability of rejection

Examples

bayes_min_n(true_p = .994,
            gamma = .6)

njjms/nicks documentation built on May 4, 2022, 8:10 a.m.