QK Box 2.3

This box continues with the Low et al. anesthetic example from Box 2.2

Preliminaries

packages: MKinfer, resample

library(MKinfer)
library(resample)

library(tidyverse)

use low data

low <- read_csv("../data/lowco2.csv")

Rows: 23 Columns: 2── Column specification ───────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
Delimiter: ","
chr (1): anesth
dbl (1): co2
ℹ Use `spec()` to retrieve the full column specification for this data.
ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

Get jackknife SE for two groups

low1 <- subset(low,anesth=="iso")
jackknife(low1$co2,mean)

Call:
jackknife(data = low1$co2, statistic = mean)
Replications: 12

Summary Statistics:

low2 <- subset(low,anesth=="ac")
jackknife(low2$co2,mean)

Call:
jackknife(data = low2$co2, statistic = mean)
Replications: 11

Summary Statistics:

Get bootstrap SE and 95%CI

low1boot <- bootstrap(low1$co2,mean,R=9999)
low1boot

Call:
bootstrap(data = low1$co2, statistic = mean, R = 9999)
Replications: 9999

Summary Statistics:

CI.percentile(low1boot, probs=c(0.025,0.975))

         2.5%    97.5%
mean 43.41667 58.16667

CI.bca(low1boot, probs=c(0.025,0.975))

         2.5% 97.5%
mean 44.41667    60

low2boot <- bootstrap(low2$co2,mean,R=9999)
low2boot

Call:
bootstrap(data = low2$co2, statistic = mean, R = 9999)
Replications: 9999

Summary Statistics:

CI.percentile(low2boot, probs=c(0.025,0.975))

     2.5%    97.5%
mean   59 86.07054

CI.bca(low2boot, probs=c(0.025,0.975))

         2.5%    97.5%
mean 60.18182 88.59399

Get bootstrap SE and CI on difference

lowboot <- bootstrap2(low$co2,mean,treatment=low$anesth,R=9999,ratio=FALSE)
lowboot

Call:
bootstrap2(data = low$co2, statistic = mean, treatment = low$anesth, 
    R = 9999, ratio = FALSE)
Replications: 9999
Two samples, sample sizes are 11 12

Summary Statistics for the difference between samples 1 and 2:

CI.percentile(lowboot, probs=c(0.025,0.975))

                 2.5%    97.5%
mean: ac-iso 6.050547 36.82406

Randomization test

perm.t.test(co2~anesth, data=low, R=9999, paired= FALSE)

    Permutation Welch Two Sample t-test

data:  co2 by anesth
(Monte-Carlo) permutation p-value = 0.0042 
permutation difference of means (SE) = 20.80499 (7.960466) 
95 percent (Monte-Carlo) permutation percentile confidence interval:
  5.575758 36.251136

Results without permutation:
t = 3.0206, df = 15.485, p-value = 0.008362
alternative hypothesis: true difference in means is not equal to 0
95 percent confidence interval:
  6.194866 35.623316
sample estimates:
 mean in group ac mean in group iso 
         70.90909          50.00000