Contrast families
Functions with an extension of .emmc provide for named contrast
families. One of the standard ones documented here may be used, or the user
may write such a function.
pairwise.emmc(levs, exclude = integer(0), include, ...) revpairwise.emmc(levs, exclude = integer(0), include, ...) tukey.emmc(levs, reverse = FALSE, ...) poly.emmc(levs, max.degree = min(6, k - 1), ...) trt.vs.ctrl.emmc(levs, ref = 1, reverse = FALSE, exclude = integer(0), include, ...) trt.vs.ctrl1.emmc(levs, ref = 1, ...) trt.vs.ctrlk.emmc(levs, ref = length(levs), ...) dunnett.emmc(levs, ref = 1, ...) eff.emmc(levs, exclude = integer(0), include, ...) del.eff.emmc(levs, exclude = integer(0), include, ...) consec.emmc(levs, reverse = FALSE, exclude = integer(0), include, ...) mean_chg.emmc(levs, reverse = FALSE, exclude = integer(0), include, ...) identity.emmc(levs, exclude = integer(0), include, ...)
levs |
Vector of factor levels |
exclude |
integer vector of indices, or character vector of levels to
exclude from consideration. These levels will receive weight 0 in all
contrasts. Character levels must exactly match elements of |
include |
integer or character vector of levels to include (the
complement of |
... |
Additional arguments, passed to related methods as appropriate |
reverse |
Logical value to determine the direction of comparisons |
max.degree |
Integer specifying the maximum degree of polynomial contrasts |
ref |
Integer(s) or character(s) specifying which level(s) to use
as the reference. Character values must exactly match elements of |
Each standard contrast family has a default multiple-testing adjustment as
noted below. These adjustments are often only approximate; for a more
exacting adjustment, use the interfaces provided to glht in the
multcomp package.
pairwise.emmc, revpairwise.emmc, and tukey.emmc generate
contrasts for all pairwise comparisons among estimated marginal means at the
levels in levs. The distinction is in which direction they are subtracted.
For factor levels A, B, C, D, pairwise.emmc generates the comparisons
A-B, A-C, A-D, B-C, B-D, and C-D, whereas revpairwise.emmc generates
B-A, C-A, C-B, D-A, D-B, and D-C. tukey.emmc invokes
pairwise.emmc or revpairwise.emmc depending on reverse.
The default multiplicity adjustment method is "tukey", which is only
approximate when the standard errors differ.
poly.emmc generates orthogonal polynomial contrasts, assuming
equally-spaced factor levels. These are derived from the
poly function, but an ad hoc algorithm is used to
scale them to integer coefficients that are (usually) the same as in
published tables of orthogonal polynomial contrasts. The default multiplicity
adjustment method is "none".
trt.vs.ctrl.emmc and its relatives generate contrasts for comparing
one level (or the average over specified levels) with each of the other
levels. The argument ref should be the index(es) (not the labels) of
the reference level(s). trt.vs.ctrl1.emmc is the same as
trt.vs.ctrl.emmc with a reference value of 1, and
trt.vs.ctrlk.emmc is the same as trt.vs.ctrl with a reference
value of length(levs). dunnett.emmc is the same as
trt.vs.ctrl. The default multiplicity adjustment method is
"dunnettx", a close approximation to the Dunnett adjustment.
Note in all of these functions, it is illegal to have any overlap
between the ref levels and the exclude levels. If any is found,
an error is thrown.
consec.emmc and mean_chg.emmc are useful for contrasting
treatments that occur in sequence. For a factor with levels A, B, C, D, E,
consec.emmc generates the comparisons B-A, C-B, and D-C, while
mean_chg.emmc generates the contrasts (B+C+D)/3 - A, (C+D)/2 -
(A+B)/2, and D - (A+B+C)/3. With reverse = TRUE, these differences go
in the opposite direction.
eff.emmc and del.eff.emmc generate contrasts that compare each
level with the average over all levels (in eff.emmc) or over all other
levels (in del.eff.emmc). These differ only in how they are scaled.
For a set of k EMMs, del.eff.emmc gives weight 1 to one EMM and weight
-1/(k-1) to the others, while eff.emmc gives weights (k-1)/k and -1/k
respectively, as in subtracting the overall EMM from each EMM. The default
multiplicity adjustment method is "fdr". This is a Bonferroni-based
method and is slightly conservative; see p.adjust.
identity.emmc simply returns the identity matrix (as a data frame),
minus any columns specified in exclude. It is potentially useful in
cases where a contrast function must be specified, but none is desired.
A data.frame, each column containing contrast coefficients for levs. The "desc" attribute is used to label the results in emmeans, and the "adjust" attribute gives the default adjustment method for multiplicity.
Caution is needed in cases where the user alters the ordering of
results (e.g., using the the "[...]" operator), because the
contrasts generated depend on the order of the levels provided. For
example, suppose trt.vs.ctrl1 contrasts are applied to two by
groups with levels ordered (Ctrl, T1, T2) and (T1, T2, Ctrl) respectively,
then the contrasts generated will be for (T1 - Ctrl, T2 - Ctrl) in the
first group and (T2 - T1, Ctrl - T1) in the second group, because the first
level in each group is used as the reference level.
warp.lm <- lm(breaks ~ wool*tension, data = warpbreaks)
warp.emm <- emmeans(warp.lm, ~ tension | wool)
contrast(warp.emm, "poly")
contrast(warp.emm, "trt.vs.ctrl", ref = "M")
# Compare only low and high tensions
# Note pairs(emm, ...) calls contrast(emm, "pairwise", ...)
pairs(warp.emm, exclude = 2)
# (same results using exclude = "M" or include = c("L","H") or include = c(1,3))
### Setting up a custom contrast function
helmert.emmc <- function(levs, ...) {
M <- as.data.frame(contr.helmert(levs))
names(M) <- paste(levs[-1],"vs earlier")
attr(M, "desc") <- "Helmert contrasts"
M
}
contrast(warp.emm, "helmert")
## Not run:
# See what is used for polynomial contrasts with 6 levels
emmeans:::poly.emmc(1:6)
## End(Not run)Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.