MeanCIn: Sample Size for a Given Width of a Confidence Interval for a Mean
Description
Returns the required sample size to obtain a given width of a confidence interval for the sample mean. The function uses uniroot()
to find a numeric solution. The t distribution is used.
Usage
MeanCIn(ci, sd, interval = c(2, 100000), conf.level = 0.95,
norm = FALSE, tol = .Machine$double.eps^0.5)
Arguments
- ci
the left and right bound of the interval, which is presumed to be symmetric.
- sd
the standard deviation of the sample.
- interval
the interval for the sample size to be searched into, (default is c(2, 100000)).
- conf.level
confidence level, defaults to 0.95
.
- norm
logical, determining if the t- or normaldistribution should be used.
- tol
the desired accuracy (convergence tolerance).
Author
Andri Signorell <andri@signorell.net>
Details
The required sample sizes for a specific width of confidence interval for the mean depends recursively on the sample size, as the samplesize defines the degrees of freedom in the t-distribution. Although in most practical cases it will be sufficient to use the normal distribution, we might be interested in exact results.