[R] Bootstrap CIs for weighted means of paired differences

[ivan]

Hi,

I am trying to compute bootstrap confidence intervals for weighted means of
paired differences with the boot package. Unfortunately, the weighted mean
estimate lies out of the confidence bounds and hence I am obviously doing
something wrong.

Appreciate any help. Thanks. Here is a reproducible example:


library(boot)
set.seed(1111)
x <- rnorm(50)
y <- rnorm(50)
weights <- runif(50)
weights <- weights / sum(weights)
dataset <- cbind(x,y,weights)
vw_m_diff <- function(dataset,w, d) {
    differences <- dataset[d,1]-dataset[d,2]
    weights <- w[d]
    return(weighted.mean(x=differences, w=weights))
}
res_boot <- boot(dataset, statistic=vw_m_diff, R = 1000, w=dataset[,3])
boot.ci(res_boot)

*BOOTSTRAP CONFIDENCE INTERVAL CALCULATIONS*
*Based on 1000 bootstrap replicates*

*CALL : *
*boot.ci <http://boot.ci>(boot.out = res_boot)*

*Intervals : *
*Level      Normal              Basic         *
*95%   (-0.8365, -0.3463 )   (-0.8311, -0.3441 )  *

*Level     Percentile            BCa          *
*95%   (-0.3276,  0.1594 )   (-0.4781, -0.3477 )  *

weighted.mean(x=dataset[,1]-dataset[,2], w=dataset[,3])

*[1] -0.07321734*

	[[alternative HTML version deleted]]