qreference {DAAG}

Simulate QQ reference plots

Description

This function computes the QQ plot for given data and specified distribution, then repeating the comparison for data simulated from the specified distribution. The plots for simulated data give an indication of the range of variation that is to expected, and thus calibrate the eye.

Usage

qreference(test = NULL, m = 30, nrep = 6, pch=c(16,2), distribution = function(x) qnorm(x, 
    mean = ifelse(is.null(test), 0, mean(test)), sd = ifelse(is.null(test), 
    1, sd(test))), seed = NULL, nrows = NULL, cex.strip = 0.75, 
    xlab = NULL, ylab = NULL) 

Arguments

test

a vector containing a sample to be tested; if not supplied, all qq-plots are for data simulated from the reference distribution

m

the sample size for the reference samples; default is test sample size if test sample is supplied

nrep

the total number of samples, including reference samples and test sample if any

pch

plot character(s)

distribution

reference distribution; default is standard normal

seed

the random number generator seed

nrows

number of rows in the plot layout

cex.strip

character expansion factor for labels

xlab

label for x-axis

ylab

label for y-axis

Value

QQ plots of the sample (if test is non-null) and all reference samples

Author(s)

J.H. Maindonald

Examples

# qreference(rt(30,4))

# qreference(rt(30,4), distribution=function(x) qt(x, df=4))

# qreference(rexp(30), nrep = 4)

# toycars.lm <- lm(distance ~ angle + factor(car), data = toycars)
# qreference(residuals(toycars.lm), nrep = 9)