2

I'm trying to repeatedly sample two vectors, x and y, 100 times and compute a t-test after each sample. I then want to combine the results into a dataframe using the broom package. My attempt at doing this is below. However, all I've managed to do is carry out 100 t-tests on the same two vectors.

library(plyr); library(broom)

x <- rnorm(10000, 3, 3)
y <- rnorm(10000, 5, 3)


x_sample <- sample(x, size = 20, replace = FALSE)
y_sample <- sample(y, size = 20, replace = FALSE)

ldply(1:100, function(x) tidy(t.test(x_sample, y_sample)))
luciano
  • 13,158
  • 36
  • 90
  • 130

1 Answers1

1

We can use replicate with mapply. We replicate the sample for n times for 'x' and 'y', and do the t.test for the corresponding list elements using mapply and then tidy the output. 

n <- 100
mapply(function(x,y) tidy(t.test(x,y)), 
       replicate(n, sample(x, size=20, replace=FALSE), simplify=FALSE), 
       replicate(n, sample(y, size=20, replace=FALSE), simplify=FALSE))
akrun
  • 874,273
  • 37
  • 540
  • 662