CODEWITHSUNDEEP
rdplyrapplysapply
Luc
Luc

Reputation: 1

How to calculate the t-test columnwise, by fixing my own true value of the mean

I'm trying to comput t-test columunwise and fixe the true value of mean (mu) to 0.9. The follwing is my code chunk.

data set

vec1 <- c(.5,.1, .03, .2, .09, .02,.3,.4,.6,.8)
vec2 <- c(5, 4,3, 8, 3, 1, 4,1, 10, 5)
vec3 <- round(rnorm(10), digits = 2)
tab <- cbind(vec1, vec2, vec3)
tab

code

I used the following apply function

k <- apply(tab, 2, t.test)
k

But this do not allow to fixe the mu. It computs the t-test with the default mu. Then, I tried to create a loop like this:

for (i in 1:ncol(tab)) {
  h=list()
  h[[i]]=t.test(tab[,i], mu=0.9)
print(h)
}

I got a list of length 3 as I want. But something strange is happening in the output.

h [[1]]

One Sample t-test

data: tab[, i] t = -7.0941, df = 9, p-value = 5.703e-05 alternative hypothesis: true mean is not equal to 0.9 95 percent confidence interval: 0.1139489 0.4940511 sample estimates: mean of x 0.304

[[1]] NULL

[[2]]

One Sample t-test

data: tab[, i] t = 3.9023, df = 9, p-value = 0.003607 alternative hypothesis: true mean is not equal to 0.9 95 percent confidence interval: 2.371053 6.428947 sample estimates: mean of x 4.4

[[1]] NULL

[2]] NULL

[[3]]

One Sample t-test

data: tab[, i] t = -2.6103, df = 9, p-value = 0.02826 alternative hypothesis: true mean is not equal to 0.9 95 percent confidence interval: -0.5485048 0.7965048 sample estimates: mean of x 0.124

With is output, when I aske for the first or second element from the list like this :

h[[1]] 
h[[2]]

I get

NULL as result.

What is going wrong in my code?

Upvotes: 0

Views: 40

Answers (1)

danlooo
danlooo

Reputation: 10637

You can use partial evaluation of the function t.test to overwrite default arguments like mu:

vec1 <- c(.5,.1, .03, .2, .09, .02,.3,.4,.6,.8)
vec2 <- c(5, 4,3, 8, 3, 1, 4,1, 10, 5)
vec3 <- round(rnorm(10), digits = 2)
tab <- cbind(vec1, vec2, vec3)
tab
#>       vec1 vec2  vec3
#>  [1,] 0.50    5  0.84
#>  [2,] 0.10    4  1.48
#>  [3,] 0.03    3 -0.93
#>  [4,] 0.20    8 -0.43
#>  [5,] 0.09    3 -2.91
#>  [6,] 0.02    1  0.33
#>  [7,] 0.30    4 -0.90
#>  [8,] 0.40    1 -0.20
#>  [9,] 0.60   10  0.75
#> [10,] 0.80    5  0.74

apply(tab, 2, purrr::partial(t.test, mu = 0.9))
#> $vec1
#> 
#>  One Sample t-test
#> 
#> data:  newX[, i]
#> t = -7.0941, df = 9, p-value = 5.703e-05
#> alternative hypothesis: true mean is not equal to 0.9
#> 95 percent confidence interval:
#>  0.1139489 0.4940511
#> sample estimates:
#> mean of x 
#>     0.304 
#> 
#> 
#> $vec2
#> 
#>  One Sample t-test
#> 
#> data:  newX[, i]
#> t = 3.9023, df = 9, p-value = 0.003607
#> alternative hypothesis: true mean is not equal to 0.9
#> 95 percent confidence interval:
#>  2.371053 6.428947
#> sample estimates:
#> mean of x 
#>       4.4 
#> 
#> 
#> $vec3
#> 
#>  One Sample t-test
#> 
#> data:  newX[, i]
#> t = -2.5636, df = 9, p-value = 0.03051
#> alternative hypothesis: true mean is not equal to 0.9
#> 95 percent confidence interval:
#>  -1.0257116  0.7797116
#> sample estimates:
#> mean of x 
#>    -0.123

Created on 2021-11-09 by the reprex package (v2.0.1)

Upvotes: 0

Related Questions