Plot Numeric (e.g. Residuals) vs 2 Factors Using Boxplots

Plots a numeric “residual like” variable against two factor covariates, using boxplots.

Usage

p.res.2fact(x, y, z, restricted, notch = FALSE,
            xlab = NULL, ylab = NULL, main = NULL)

Arguments

x,y: two factors or numeric vectors giving the levels of factors.
z: numeric vector of same length as x and y, typically residuals.
restricted: positive value which truncates the size. The corresponding symbols are marked by stars.
notch: logical indicating if the boxplots should be notched, see boxplot(*,notch).
xlab,ylab: axis labels, see plot.default, per default the actual argument expressions.
main: main title passed to plot, defaulting to the deparsed z argument.

Details

if values are restricted, this make use of the auxiliar function u.boxplot.x.

Author

Lorenz Gygax logyg@wild.unizh.ch and Martin Maechler, Jan.95; starting from p.res.2x().

Examples

I <- 8; J <- 3; K <- 20
xx <- factor(rep(rep(1:I, rep(K,I)),J))
yy <- factor(rep(1:J, rep(I*K,J)))
zz <- rt(I*J*K, df=5) #-- Student t with 5 d.f.
p.res.2fact(xx,yy,zz, restr= 4, main= "i.i.d. t <- 5 random  |.| <= 4")
#> ..Cut z= 10.75 -6.52  at  xx = 5 7 ,   yy = 3 
#> ..Cut z= 5.31 4.01  at  xx = 1 5 ,   yy = 2 
#> ..Cut z= -4.39 -4.48  at  xx = 2 8 ,   yy = 1 
#> 6 restricted observation(s)
mtext("p.res.2fact(xx,yy,zz, restr= 4, ..)",
      line=1, adj=1, outer=TRUE, cex=1)


## Real data
data(warpbreaks)
(fm1 <- lm(breaks ~ wool*tension, data = warpbreaks))
#> 
#> Call:
#> lm(formula = breaks ~ wool * tension, data = warpbreaks)
#> 
#> Coefficients:
#>    (Intercept)           woolB        tensionM        tensionH  woolB:tensionM  
#>           44.6           -16.3           -20.6           -20.0            21.1  
#> woolB:tensionH  
#>           10.6  
#> 
## call via formula method of p.res.2x():
p.res.2x(~ ., fm1) # is shorter than, but equivalent to

## p.res.2x(~ wool + tension, fm1)  ##  or the direct
## with(warpbreaks, p.res.2fact(wool, tension, residuals(fm1)))
##
## whereas this is "transposed":
p.res.2x(~ tension+wool, fm1)