CIsim: Simulated confidence intervals

In fastR: Data sets and utilities for Foundations and Applications of Statistics by R Pruim

Description

Confidence intervals calculated from simulated data are stored in a data frame that allows further processing (including plotting).

Usage

CIsim(n, samples = 100, rdist = rnorm, args = list(), 
	estimand = 0, conf.level = 0.95, 
	method = t.test, method.args = list(), 
	interval = function(x) {
    	do.call(method, c(list(x, conf.level = conf.level), method.args))$conf.int
	}, 
	estimate = function(x) {
    	do.call(method, c(list(x, conf.level = conf.level), method.args))$estimate
	}, 
	verbose = TRUE)

Arguments

n	sample size
samples	number of samples to simulate
rdist	random distribution function for generating random samples; defaults to rnorm
args	arguments for rdist
estimand	value of estimand
conf.level	level of confidence
method	function for computing a confidence interval and point estimate from data. method should return a list or structure containing named components conf.int and estimate (as in the default t.test) unless interval and estimate are supplied.
method.args	additional arguments for method.
interval	a function for computing a confidence interval from data
estimate	a function for computing the point estimate from data
verbose	should a report of percentage of simulated intervals that cover the estimand be printed to the screen?

Details

This function is design for flexibility and ease of use rather than for efficiency. The goal is to make it easy to estimate coverage rates for a wide range of confidence intervals and population distributions. The resulting data frame stores data in a manner suitable for plotting with xYplot() (see examples).

Value

A data frame with one row per sample and variables estimate (point estimate), lower (lower bound of confidence interval), upper (upper bound of confidence interval), cover (boolean indicating if estimand is in the interval), and sample (a counter, primarily useful for plotting)

Author(s)

Randall Pruim

See Also

xYplot in Hmisc.

Examples

CIsim(n=10,100) -> foo
if (require(Hmisc)) {
	xYplot(Cbind(estimate,lower,upper)~sample, foo,groups=cover, col=c('red','navy'),lwd=3,cap=0)
}

fastR documentation built on May 2, 2019, 5:53 p.m.