n_gamma {skewsamp}

Calculate sample size for gamma distribution

Description

Estimation of required sample size as given by Cundill & Alexander (2015).

Usage

n_gamma(
  mean0,
  effect,
  shape0,
  shape1 = shape0,
  alpha = 0.05,
  power = 0.9,
  q = 0.5,
  link = c("log", "identity"),
  two_sided = TRUE
)

Arguments

mean0

Mean in control group

effect

Effect size, 1 - (\mu_1 / \mu_0), where \mu_0 is the mean in the control group (mean0) and \mu_1 is the mean in the treatment group.

shape0

Shape parameter in control group

shape1

Shape parameter in treatment group. Defaults to shape0, because GLM assumes equal shape across groups.

alpha

Type I error rate

power

1 - Type II error rate

q

Proportion of observations allocated to the control group

link

Link function to use. Currently implement: 'log' and 'identity'

two_sided

logical, if TRUE the sample size will be calculated for a two-sided test. Otherwise, the sample size will be calculated for a one-sided test.

Value

Returns an object of class "sample_size". It contains the following components:

N

the total sample size

n0

sample size in Group 0 (control group)

n1

sample size in Group 1 (treatment group)

two_sided

logical, TRUE, if the estimated sample size refers to a two-sided test

alpha

type I error rate used in sample size estimation

power

target power used in sample size estimation

effect

effect size used in sample size estimation

effect_type

short description of the type of effect size

comment

additional comment, if there is any

call

the matched call.

References

Cundill, B., & Alexander, N. D. E. (2015). Sample size calculations for skewed distributions. BMC Medical Research Methodology, 15(1), 1–9. https://doi.org/10.1186/s12874-015-0023-0

Examples

n_gamma(mean0 = 8.46, effect = 0.7, shape0 = 0.639,
           alpha = 0.05, power = 0.9)