power_analysis_binomial: Calculate minimum sample size needed for binomial experiment

In scottfrechette/funcyfrech: Frechette Functions

Calculate minimum sample size needed for binomial experiment

Description

Either null_value or hdi_max_width must be null, depending whether the goal is reject a particular null value or estimating parameter with specified precision

Usage

power_analysis_binomial(
  gen_prior_mode,
  gen_prior_n,
  hdi_max_width = NULL,
  null_value = NULL,
  rope = c(max(0, null_value - 0.02), min(1, null_value + 0.02)),
  desired_power = 0.8,
  aud_prior_mode = 0.5,
  aud_prior_n = 2,
  hdi_mass = 0.95,
  init_samp_size = 20,
  verbose = TRUE
)

Arguments

gen_prior_mode	Expected mode of prior
gen_prior_n	N of prior (higher leads to more precision)
hdi_max_width	Maximum width of HDI
null_value	Null value
rope	Region of practical equivalence around parameter
desired_power	Level of power
aud_prior_mode	Prior understanding of mode from audience
hdi_mass	Range of HDI
init_samp_size	Initial sample size
verbose	Should each attempt be displayed?
aud_prior	Prior understanding of variation from audience

Details

Function was derived by Solomon Kurz's bookdown converting Kruschke's book into brms and tidyverse

Source

https://bookdown.org/content/3686/goals-power-and-sample-size.html#computing-power-and-sample-size/

scottfrechette/funcyfrech documentation built on Aug. 26, 2022, 9:13 a.m.