ciplot: Effect plot with multiple CIs for different groups

In broman: Karl Broman's R Code

Effect plot with multiple CIs for different groups

Description

Uses grayplot() to plot a set of confidence intervals.

Usage

ciplot(
  est,
  se = NULL,
  lo = NULL,
  hi = NULL,
  SEmult = 2,
  labels = NULL,
  rotate = FALSE,
  ...
)

Arguments

est	Vector of estimates
se	Vector of standard errors
lo	Vector of lower values for the intervals
hi	Vector of upper values for the intervals
SEmult	SE multiplier to create intervals
labels	Labels for the groups (vector of character strings)
rotate	If TRUE, have group as y-axis; default (FALSE) has group on x-axis.
...	Optional graphics arguments

Details

Calls grayplot() with special choices of graphics parameters, as in dotplot().

Provide either se or both lo and hi. In the case that se is used, the intervals will be est +/- SEmult * se.

If labels is not provided, group names are taken from the names(est). If that is also missing, we use capital letters.

You can control the CI line widths with ci_lwd and the color of the CI segments with ci_col. You can control the width of the segments at the top and bottom with ci_endseg.

Value

None.

See Also

grayplot(), dotplot()

Examples

x <- rnorm(40, c(1,3))
g <- rep(c("A", "B"), 20)
me <- tapply(x, g, mean)
se <- tapply(x, g, function(a) sd(a)/sqrt(sum(!is.na(a))))
ciplot(me, se) # default is +/- 2 SE
ciplot(me, se, SEmult=1)
ciplot(me, se, rotate=TRUE)
lo <- me - 2*se
hi <- me + 2*se
ciplot(me, lo=lo, hi=hi)

broman documentation built on May 29, 2024, 7:18 a.m.