RVAideMemoire (version 0.9-29-2)

mat.cont: Matrix of contrasts generation

Description

Return a matrix of contrasts for pairwise comparisons. By default all pairwise comparisons are returned.

Usage

mat.cont(lev, ref = NULL, restrict = NULL, exclude = FALSE)

Arguments

lev

a vector of character strings giving levels to be compared. This includes levels of a single factor or levels of an interaction (between factors or factors and covariates).

ref

optional reference level (only contrasts involving this level are generated).

restrict

optional vector of levels (only contrasts involving these levels are generated).

exclude

logical. If TRUE and if restrict is not NULL, levels which are not involved in comparisons are not present in the matrix (to be used when the model has been generated with the subset option).

Details

In these matrices, each line is a comparison (= contrast) and each colum is a level of the factor. Rules for writing contrasts are: - levels not involved in the comparison must have a null value - levels to be compared must have opposite signs - levels can be grouped (for example 2 -1 -1 give a comparison of the first level against the group composed by the two others) - the sum of all values of a contrast must be null.

See Also

lsmeans, user.cont

Examples

# A factor with 5 levels
fact <- gl(5,10,labels=LETTERS[1:5])

# A matrix of contrasts with all possible pairwise comparisons
mat.cont(levels(fact))

# A matrix with only comparisons to level "A"
mat.cont(levels(fact),ref="A")

# A matrix with all pairwise comparisons between "A", "B" and "E"
mat.cont(levels(fact),restrict=c("A","B","E"))

# The same with levels "C" and "D" excluded
mat.cont(levels(fact),restrict=c("A","B","E"),exclude=TRUE)


# With an interaction between two factors
fact2 <- gl(2,25,labels=letters[1:2])
mat.cont(levels(fact:fact2))


# With an interaction between a factor and a covariate
covar <- runif(50,10,20)
lev <- paste("covar:fact",levels(fact),sep="")
mat.cont(lev)