1
$\begingroup$

If I want to determine the difference between two correlated measurements (i.e., pre-post scores), is a one-sample t-test on the difference scores compared to 0 an identical test to a paired sample t-test?

$\endgroup$
  • $\begingroup$ compare the formulas.... $\endgroup$ – Glen_b Nov 22 '17 at 1:38
1
$\begingroup$

Yes. We can test this easily enough in R or whatever you like e.g. (R code):

set.seed(123)

x1 <- rnorm(100)
x2 <- rnorm(100)
diff <- x1-x2
t.test(x1,x2, paired = TRUE)
t.test(diff)

but it also follows from the definition of a paired t-test.

$\endgroup$
  • $\begingroup$ More relevant is your point that it follows by the definition. There is nothing magic about two functions in R that do exactly the same computations. $\endgroup$ – Michael Chernick Nov 22 '17 at 5:18
  • $\begingroup$ They do the same computations because of the definitions and because R implemented the definitions. $\endgroup$ – Peter Flom - Reinstate Monica Nov 22 '17 at 13:06

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.