designr (version 0.1.13)

design.contrasts: Retrieve contrast codes for a design

Description

This function can be used to retrieve contrast codes based on experimental codes / planned observations.

Use this function to retrieve the names for contrasts in an experimental design, such as used by lm, lmer and many other regression models.

Usage

design.contrasts(
  design,
  factors = names(fixed.factors(design)),
  contrasts = NULL,
  expand = TRUE,
  rename_contrasts = "%1$s%2$s",
  intercept = FALSE,
  interactions = FALSE,
  include_random_levels = FALSE
)
contrast.names(design, ranfac = NULL, as_symbols = FALSE, ...)

Value

A design matrix (if expand==TRUE, default) or a list of factor levels (if expand==FALSE) for design.contrasts or contrast names for contrast.names.

Arguments

design

A design object

factors

Which fixed factors to include in the output

contrasts

Contrasts to use for each categorical factor. Should be a list named after the fixed effects and containing contrast matrices, such as the ones generated by the standard contrast functions. If NULL or the fixed factor is not found in the list, default contrasts are used (typically, treatment contrasts).

expand

If TRUE, a design matrix is returned. If FALSE, all factors (and interactions) with their respective levels are returned.

rename_contrasts

This is the pattern after which columns in the design matrix are named. By default, this is a direct concatenation of factor name and contrast name.

intercept

If TRUE, an intercept is added to the matrix. Its value is 1 for all observations.

interactions

If TRUE, interactions of fixed factors are included.

include_random_levels

If TRUE, levels of random factors are included in the matrix.

ranfac

Return random-effects contrast names for a given random factor. If NULL, return fixed-effects contrast names.

as_symbols

Return contrast names as symbols rather than strings (character vectors).

...

Arguments to pass on to design.contrasts()

Functions

• contrast.names(): Retrieve contrast names for a design

Examples

des <- fixed.factor("Factor1", c("1A","1B")) +
       fixed.factor("Factor2", c("2A","2B")) +
       random.factor("Subject", c("Factor1"))

              
design.contrasts(des)

contrast.names(des)

stopifnot(contrast.names(des) == c("Factor11B", "Factor22B"))

contrast.names(des, as_symbols = TRUE)

design.contrasts(des, contrasts = list(Factor2 = contr.sum))

contrast.names(des, contrasts = list(Factor2 = contr.sum))