contrMatRatio {mratios} | R Documentation |
Creates numerator and denominator contrast matrices for ratio-based hypotheses for common multiple comparison and trend test problems
Description
Creates numerator and denominator contrast matrices for some common multiple comparison and trend test problems. These matrices are internally used by the sci.ratio and simtest.ratio functions. The contrMatRatio function is a modification of the function contrMat (multcomp).
Whether the given definitions of contrast matrices for trend test problems in terms of ratios make sense and how they are to be interpreted is to be discussed.
Usage
contrMatRatio(n, type = c("Tukey", "Dunnett", "Sequen",
"AVE", "GrandMean", "Changepoint", "Marcus", "McDermott",
"Williams", "UmbrellaWilliams"), base = 1)
Arguments
n |
integer vector of sample sizes |
type |
the type of multiple contrasts
|
base |
a single integer specifying the control (i.e. denominator) group for "Dunnett"-type contrasts for calculating the ratios to the control |
Details
This is a simple adaption of the contrMat function in the package multcomp for ratio hypotheses.
Value
A list containing:
numC |
the (named) numerator contrast where rows correspond to contrasts |
denC |
the (named) denominator contrast where rows correspond to contrasts |
rnames |
a character vector with names of the contrasts |
and the type of contrast as attr.
Author(s)
Frank Schaarschmidt and Daniel Gerhard by modifying the code of contrMat(multcomp)
See Also
contrMat(multcomp)
Examples
library(mratios)
n=c(A=10,B=20,Z=10,D=10)
contrMatRatio(n=n, type="Dunnett", base=1)
contrMatRatio(n=n, type="Dunnett", base=3)
contrMatRatio(n=n, type="Tukey")
contrMatRatio(n=n, type="Sequen")
contrMatRatio(n=n, type="AVE")
contrMatRatio(n=n, type="GrandMean")
contrMatRatio(n=n, type="Williams")
contrMatRatio(n=n, type="UmbrellaWilliams")