m-py / prmisc Goto Github PK

When printing a one-sample t-test + effect size estimate using print_ttest(), the following error is thrown: t-test is not paired, but Cohens d is paired

See:

library(effectsize)
library(prmisc)

my_data <- rnorm(100, .2)
my_ttest <- t.test(my_data)
my_cohen <- cohens_d(my_data)

print_ttest(my_ttest, my_cohen)

That is because print_ttest() checks whether the attribute "paired" for the Cohens d object is FALSE. However, the attribute is TRUE for single samples.

My suggestion would be to adjust the check for the t-test object so it also covers the single sample case, but wanted to discuss first whether that would be desired behaviour.

print_ttest should compute effect size from the t-test object instead of requiring an object from the effsize package.