Skip to contents

Confidence Interval For Difference of Means

Source: R/meanDiffCI.R
meanDiffCI.Rd

Calculates the confidence interval for the difference of two means either the classical way or with the bootstrap approach.

Usage

meanDiffCI(
  x,
  y,
  conf.level = 0.95,
  sides = c("two.sided", "left", "right"),
  method = c("classic", "boot"),
  paired = FALSE,
  var.equal = FALSE,
  na.rm = FALSE,
  ...
)

Arguments

x

a (non-empty) numeric vector of data values.

y

a (non-empty) numeric vector of data values.

conf.level

confidence level of the interval.

sides

a character string specifying the side of the confidence interval, must be one of "two.sided" (default), "left" or "right". You can specify just the initial letter. "left" would be analogue to a hypothesis of "greater" in a t.test.

method

a vector of character strings representing the type of intervals required. The value should be any subset of the values "classic", "boot". Bootstrap type can be provided by the dots. See boot::boot.ci().

paired

a logical indicating whether you want confidence intervals for a paired design. Defaults to FALSE.

var.equal

a logical variable indicating whether to treat the two variances as being equal. Default is FALSE. If TRUE then the pooled variance is used to estimate the variance otherwise the Welch (or Satterthwaite) approximation to the degrees of freedom is used. Passed on to t.test().

na.rm

logical. Should missing values be removed? Defaults to FALSE.

...

further arguments, can be used to provide further arguments to the boot function.

Value

A named numeric vector with elements:

meandiff

point estimate, the difference: mean(x) - mean(y)

lci

lower confidence interval bound

uci

upper confidence interval bound

Details

This function collects code from two sources. The classical confidence interval is calculated by means of t.test(). The bootstrap intervals are strongly based on the example in boot::boot().

The bootstrap type "stud" (studentized) is not supported: the statistic functions used here return only the point estimate, not a per-replicate variance estimate, so requesting it raises an error.

See also

varCI(), boot::boot.ci()

Other ci.location: meanCI(), meanCIn(), medianCI(), quantileCI(), sumCI()

Examples


x <- mtcars[mtcars$am == 0, "mpg"]
y <- mtcars[mtcars$am == 1, "mpg"]

meanDiffCI(x, y, na.rm=TRUE)
#>   meandiff        lci        uci 
#>  -7.244939 -11.280194  -3.209684 
meanDiffCI(x, y, conf.level=0.99, na.rm=TRUE)
#>   meandiff        lci        uci 
#>  -7.244939 -12.769128  -1.720751 

# the different types of bootstrap confints
meanDiffCI(x, y, method="boot", type="norm", na.rm=TRUE)
#>   meandiff        lci        uci 
#>  -7.244939 -10.852192  -3.676966 
meanDiffCI(x, y, method="boot", type="basic", na.rm=TRUE)
#>   meandiff        lci        uci 
#>  -7.244939 -10.595951  -3.658704 
# type="stud" is not supported (see Details) and raises an error:
# meanDiffCI(x, y, method="boot", type="stud", na.rm=TRUE)
meanDiffCI(x, y, method="boot", type="perc", na.rm=TRUE)
#>   meandiff        lci        uci 
#>  -7.244939 -11.010121  -3.540486 
meanDiffCI(x, y, method="boot", type="bca", na.rm=TRUE)
#>   meandiff        lci        uci 
#>  -7.244939 -10.929270  -3.724819 

# for long form variables
with(mtcars, with(split(mpg, am), 
  meanDiffCI(`0`, `1`) )
)
#>   meandiff        lci        uci 
#>  -7.244939 -11.280194  -3.209684