mCI: Interactive applets to explore confidence interval coverage

Description Usage Arguments Details Value Author(s) Examples

Description

These applets allow the user explore the coverage of confidence intervals computed from many random samples.

Usage

1
2

Arguments

...

additional arguments

Details

Data for mCIprop is created using rbinom. The user may choose to calculate confidence intervals using a Wald interval or an Agresti interval by using the type of CI picker. There are sliders in the manipulate box to interact with sample size, confidence level, true mean, number of trials, and random seed. Number of trials increases the number of lines plotted. The total number of confidence intervals not containing the true mean, or "failed CIs", is printed above the plot.

Known bugs: Color scheme may be redone, orange/chocolate1 may not be the best color for failed CIs. Wald CIs "error" when true mean is near the edge of the line and sample size is low, however this is more a product of the Wald CI calculation than the program.

Value

A function that allows the user to explore confidence interval coverage interactively.

Author(s)

Andrew Rich (andrew.joseph.rich@gmail.com) and Daniel Kaplan (kaplan@macalester.edu)

Examples

1
2
3
4
	if (require(manipulate)) {
		mCIprop()
		mCIt()
	}

mosaicManip documentation built on May 2, 2019, 5:48 p.m.