Description Usage Arguments Value Author(s) See Also Examples
This function performs all pairwise compararisons among
means returning pontual and intervalar estimates followed by
letters to easy discriminate values. It is in fact a wraper of
glht()
.
1 |
X |
a matrix where each line is a linear function of the model
parameters to estimate a least squares mean. In most pratical
cases, it is an object from the |
model |
a model with class recognized by
|
focus |
a string with the name of the factor which levels are being compared. |
test |
a p-value correction method. See
|
level |
the experimentwise significance level for the multiple
comparisons. The individual coverage of the confidence interval
is |
cld2 |
Logical, if |
a data.frame
with interval estimates and compact
letter display for the means comparisons.
Walmes Zeviani, walmes@ufpr.br.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 | library(doBy)
library(multcomp)
# Single factor.
m0 <- lm(weight ~ feed, data = chickwts)
anova(m0)
# Prepare the matrix to estimate lsmeans.
L <- LE_matrix(m0, effect = "feed")
rownames(L) <- levels(chickwts$feed)
apmc(L, model = m0, focus = "feed", test = "fdr")
data(warpbreaks)
# Two factors (complete factorial).
m1 <- lm(breaks ~ wool * tension, data = warpbreaks)
anova(m1)
L <- LE_matrix(m1, effect = c("wool", "tension"))
attributes(L)
Ls <- by(L, INDICES = attr(L, "grid")$tension, FUN = as.matrix)
Ls <- lapply(Ls, "rownames<-", levels(warpbreaks$wool))
# Comparing means of wool in each tension.
lapply(Ls, apmc, model = m1, focus = "wool",
test = "single-step", level = 0.1)
# Two factors (incomplete factorial).
warpbreaks <- subset(warpbreaks, !(tension == "H" & wool == "A"))
xtabs(~tension + wool, data = warpbreaks)
# There is NA in the estimated parameters.
m2 <- lm(breaks ~ wool * tension, data = warpbreaks)
coef(m2)
X <- model.matrix(m2)
b <- coef(m2)
X <- X[, !is.na(b)]
# unique(X)
# Uses the full estimable model matriz.
m3 <- update(m2, . ~ 0 + X)
# These models are in fact the same.
anova(m2, m3)
# LS matrix has all cells.
L <- LE_matrix(m2, effect = c("wool", "tension"))
g <- attr(L, "grid")
L <- L[, !is.na(b)]
i <- 5
L <- L[-i, ]
g <- g[-i, ]
rownames(L) <- make.names(g$tension, unique = FALSE)
Ls <- split.data.frame(L, g$wool)
# LSmeans with MCP test.
lapply(Ls, apmc, model = m3, focus = "tension",
test = "single-step", level = 0.1, cld2 = TRUE)
# Sample means.
aggregate(breaks ~ tension + wool, data = warpbreaks, FUN = mean)
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.