rbind.emmGrid: Combine or subset 'emmGrid' objects

In rvlenth/emmeans: Estimated Marginal Means, aka Least-Squares Means

Combine or subset emmGrid objects

Description

These functions provide methods for rbind and [ that may be used to combine emmGrid objects together, or to extract a subset of cases. The primary reason for doing this would be to obtain multiplicity-adjusted results for smaller or larger families of tests or confidence intervals.

Usage

## S3 method for class 'emmGrid'
rbind(..., deparse.level = 1, adjust = "bonferroni")

## S3 method for class 'emmGrid'
e1 + e2

## S3 method for class 'emmGrid'
x[i, adjust, drop.levels = TRUE, ...]

## S3 method for class 'emmGrid'
head(x, n = 6, ...)

## S3 method for class 'emmGrid'
tail(x, n = 6, ...)

## S3 method for class 'emmGrid'
subset(x, subset, ...)

## S3 method for class 'emm_list'
rbind(..., which, adjust = "bonferroni")

force_regular(object)

Arguments

...	In rbind, object(s) of class emmGrid. In others, additional arguments passed to other methods
deparse.level	(required but not used)
adjust	Character value passed to update.emmGrid
e1, e2, x, object	Objects of class emmGrid
i	Integer vector of indexes
drop.levels	Logical value. If TRUE, the "levels" slot in the returned object is updated to hold only the predictor levels that actually occur
n	integer number of entries to include (or exclude if negative)
subset	logical expression indicating which rows of the grid to keep
which	Integer vector of subset of elements to use; if missing, all are combined

Value

A revised object of class emmGrid

The result of e1 + e2 is the same as rbind(e1, e2)

The rbind method for emm_list objects simply combines the emmGrid objects comprising the first element of ....

force_regular adds extra (invisible) rows to an emmGrid object to make it a regular grid (all combinations of factors). This regular structure is needed by emmeans. An object can become irregular by, for example, subsetting rows, or by obtaining contrasts of a nested structure.

Note

rbind throws an error if there are incompatibilities in the objects' coefficients, covariance structures, etc. But they are allowed to have different factors; a missing level '.' is added to factors as needed.

These functions generally reset by.vars to NULL; so if you want to keep any “by” variables, you should follow-up with update.emmGrid.

Examples

warp.lm <- lm(breaks ~ wool * tension, data = warpbreaks)
warp.rg <- ref_grid(warp.lm)

# Do all pairwise comparisons within rows or within columns, 
# all considered as one faily of tests:
w.t <- pairs(emmeans(warp.rg, ~ wool | tension))
t.w <- pairs(emmeans(warp.rg, ~ tension | wool))
rbind(w.t, t.w, adjust = "mvt")
update(w.t + t.w, adjust = "fdr")  ## same as above except for adjustment

# Show only 3 of the 6 cases
summary(warp.rg[c(2, 4, 5)])

# After-the-fact 'at' specification
subset(warp.rg, wool == "A")  ## or warp.rg |> subset(wool == "A")


### Working with 'emm_list' objects
mod <- lm(conc ~ source + factor(percent), data = pigs)
all <- emmeans(mod, list(src = pairwise ~ source, pct = consec ~ percent))
rbind(all, which = c(2, 4), adjust = "mvt")

### Irregular object
tmp <- warp.rg[-1]
## emmeans(tmp, "tension")   # will fail because tmp is irregular
emmeans(force_regular(tmp), "tension")   # will show some results

rvlenth/emmeans documentation built on April 25, 2024, 1:13 a.m.