Power Calculation for Comparing Two Geometric Rates.
Compute sample size or power for comparing two geometric rates.
power_Geometric(n1 = NULL, n2 = NULL, mu1 = NULL, mu2 = NULL,
sig.level = 0.05, power = NULL, duration = 1,
equal.sample = TRUE, alternative = c("two.sided", "one.sided"),
approach = 3)n1 |
sample size in group 1, or sample size in each group if |
n2 |
sample size in group 2 |
mu1 |
expected rate of events per time unit for group 1 |
mu2 |
expected rate of events per time unit for group 2 |
sig.level |
significance level (Type I error probability) |
power |
power of test (1 minus Type II error probability) |
duration |
(average) treatment duration |
equal.sample |
equal sample sizes for two groups, see details |
alternative |
one- or two-sided test |
approach |
1, 2, or 3; see Zhu and Lakkis (2014). |
Exactly one of the parameters n1, n2, and power must be passed as NULL, and that parameter is determined from the others.
If equal.sample = TRUE is used, N in output will denote the number in each group.
Since the geometric distribution is a special case of negative binomial distribution, we used the algorithm for negative binomial distribution with setting theta = 1.
The computations are based on the formulas given in Zhu and Lakkis (2014).
For approach 1, the events rates under null hypothesis are set as mu1 for both group 1 and group 2.
For approach 2, the events rates under null hypothesis are set as mu1 and mu2 for group 1 and group 2, respectively.
For approach 3, the events rates under null hypothesis are set as the maximum likelihood estimation of the overall events rate for both group 1 and group 2.
Object of class "power.htest", a list of the arguments (including the computed one) augmented with note and method elements.
# calculate power, equal sizes power_Geometric(n1 = 100, mu1 = 0.3, mu2 = 0.6) # calculate power, unequal sizes power_Geometric(n1 = 180, n2 = 140, mu1 = 0.3, mu2 = 0.5) # calculate n power_Geometric(mu1 = 0.3, mu2 = 0.4, power = 0.8)
Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.