likertWeighted: Special case wrapper for likert() when multiple columns are...

In HH: Statistical Analysis and Data Display: Heiberger and Holland

Special case wrapper for likert() when multiple columns are to have the same bar thicknesses. Uses formula with one or two conditioning variables.

Description

Special case wrapper for likert() when multiple columns are to have the same bar thicknesses. Uses formula with one or two conditioning variables.

Usage

likertWeighted(x, ...) ## generic

## S3 method for class 'array'
likertWeighted(x, ..., C = 1, Q = 3, R = 2) ## array

## Default S3 method:
likertWeighted(x, ...) ## matrix, table, data.frame

## S3 method for class 'formula'
likertWeighted(x, data,
               xlim=c(-100, 100),
               scales=list(y=list(relation="free", cex=1.3),
                           x=list(at=seq(-100, 100, 50),
                           labels=abs(seq(-100, 100, 50)), cex=.5)),
               box.ratio=1000,
               as.percent=TRUE, rightAxis=FALSE,
               between=list(x=1, y=0),
               strip=FALSE, strip.left=FALSE,
               par.settings=list(clip=list(panel="off")),
               h.resizePanels=1,
               auto.key.title=NULL,
               auto.key.columns=dim(data)[[2]] -
               NumberOfConditioningVariables(formula), ## excludes conditioning variables
               auto.key.cex=1.2,
               auto.key.cex.title=1.2,
               auto.key.lines.title=3,
               ylab=NULL,
               axis.top=dimnames(result)[[1]], ## Questions
               axis.top.row=1,
               ...)
               

Arguments

x	For the default method, a matrix or data.frame or two-dimensional table. For the array method, a two- or-three-dimensional array. For the formula method, a formula.
formula	Standard trellis formula, usually ~ . | row + column or ~ . | row
data	A data.frame that has been constructed from a 2D object (matrix or table or data.frame) to include an additional column row, or constructed from a 3D array by toCQxR to include two additional columns group and row. The default and array methods do that construction.
C, R, Q	Integers, one each of 1,2,3; positions of the three dimensions. Used in array method. See toCQxR.
xlim, between, strip, strip.left, par.settings, ylab	See xyplot
scales	See xyplot. For likertWeighted, when scales for x is changed, scales for y must be stated also.
box.ratio	See panel.bwplot.
as.percent, rightAxis, ..., h.resizePanels	First see the formula method for likertWeighted, and then likert.
auto.key.title, auto.key.columns, auto.key.cex, auto.key.cex.title, auto.key.lines.title	Values which will be used in trellis argument auto.key=list(title=auto.key.title, columns=auto.key.columns, cex=auto.key.cex, cex.title=auto.key.cex.title, lines.title=auto.key.lines.title)
axis.top	Label to be placed at x=0 for top (and other specified) panel of each column.
axis.top.row	Which rows will have axis.top displayed.

Value

A likert plot as a "trellis" object.

Author(s)

Richard M. Heiberger <[e-mail]>

See Also

likert

Examples

## simplest 2D example
tmp <- matrix(1:12, 3, 4,
              dimnames=list(c("A","B","C"),
                            c(letters[4:7]))) * c(1,2,3)
tmp
rowSums(tmp)

likertWeighted(tmp,
               h.resizePanels=rowSums(tmp),
               main="likertWeighted, simplest example,
                     defaults to Percent,
                     specified row thicknesses")

## Same example with explicit use of the formula method
## (default method does this for you).
tmpdd <- data.frame(tmp, row=row.names(tmp))
tmpdd
likertWeighted(~ . | row, tmpdd,            ## tmpdd
               h.resizePanels=rowSums(tmp), ## tmp
main="likertWeighted,
same example but with explicit formula method")

## show subgroups
likertWeighted(tmp,
               h.resizePanels=rowSums(tmp),
               between=list(y=c(0, 1)),
               ylab=c("C in its own group","A and B together"),
               main="between=list(y=c(0,1) ## standard lattice between argument
Adjacent A and B with y.between = 0 are in the same bordered group.
Adjacent B and C with y.between != 0 are in different bordered groups.")


## simplest 3D example
## This is natural when multiple questions are asked of the
## same set of respondents in a survey.
## This example simulates that situation.
##
tmp3D <- abind::abind(h=tmp, i=tmp, j=tmp, along=3)
tmp3D[1,,"i"] <- tmp3D[1,c(4,2,1,3),"h"]
tmp3D[2,,"i"] <- tmp3D[2,c(2,4,3,1),"h"]
tmp3D[3,,"i"] <- tmp3D[3,c(4,1,2,3),"h"]
tmp3D[1,,"j"] <- tmp3D[1,c(4,3,2,1),"h"]
tmp3D[2,,"j"] <- tmp3D[2,c(1,4,3,2),"h"]
tmp3D[3,,"j"] <- tmp3D[3,c(2,4,3,1),"h"]
## now
rowSums(tmp3D[,,1]) == rowSums(tmp3D[,,2])
rowSums(tmp3D[,,1]) == rowSums(tmp3D[,,3])


likertWeighted(tmp3D, h.resizePanels=rowSums(tmp3D[,,1]),
               main="simplest 3D example, array method")

likertWeighted(tmp3D, h.resizePanels=rowSums(tmp3D[,,1]),
               between=list(x=1, y=c(0, 1)),
               main="simplest 3D example, array method, with subgroups")

## Same example with explicit use of the formula method
## (array method does this for you).
tmp3Ddf <- toCQxR(tmp3D)
dimnames(tmp3Ddf)
tmp3Ddf
likertWeighted(~ . | group + row, tmp3Ddf, h.resizePanels=rowSums(tmp3D[,,1]),
               main="simplest 3D example, formula method")

HH documentation built on May 29, 2024, 6:24 a.m.