
MMC (Mean--mean Multiple Comparisons) plots in lattice
mmcplot(mmc, ...)
# S3 method for mmc
mmcplot(mmc, col=c("black","red"), lwd=c(1,1), lty=c(2,1), ...,
style=c("isomeans", "confint", "both"),
type=c("mca", "lmat", "linfct", "none"))
# S3 method for glht
mmcplot(mmc, col=c("black","red"), lwd=c(1,1), lty=c(2,1),
focus=mmc$focus, ...)
# S3 method for mmc.multicomp
mmcplot(mmc, col=c("black","red"), lwd=c(1,1), lty=c(2,1), ...)
# S3 method for multicomp
mmcplot(mmc, col=c("black","red"), lwd=c(1,1), lty=c(2,1), ...)
# S3 method for default
mmcplot(mmc, ...)
mmc
object or other object as indicated by method.
Standard
xyplot
arguments applied to the line segments
representing the contrasts.
Name of the factor for which the glht
object was constructed.
Other arguments to be passed on to the functions called by the methods.
Style of graph: The default isomeans
is the standard MMC plot with the
isomeans grid. confint
is a confidence interval plot, similar to
the plot produced by multcomp:::plot.glht
.
both
prints both the isomeans
and
the confint
plot as two panels of a trellis
structure.
When the underlying sets of means are close to each other, there will
of necessity be overprinting in the isomeans
panel and the
confint
panel will be needed as a tiebreaker.
By default the xlim
for the confint
style will match the
xlim
of the corresponding isomeans
plot.
mca
for the default paired-comparisons plot. lmat
or
linfct
for a user-specified set of contrasts. none
for confidence
intervals on the set of group means (that is, no comparisons).
A trellis
object containing the graphs.
See mmc
for the references.
mmc
for the discussion of the MMC and for many examples.
The functions mmcisomeans
, mmcmatch
,
mmcboth
are the internal functions that do the
actual work of plotting.
# NOT RUN {
data(catalystm)
catalystm1.aov <- aov(concent ~ catalyst, data=catalystm)
catalystm.mmc <-
mmc(catalystm1.aov, linfct = mcp(catalyst = "Tukey"))
mmcplot(catalystm.mmc)
mmcplot(catalystm.mmc, style="both", MMCname="catalyst")
# }
Run the code above in your browser using DataLab