I would like to obtain a percent difference between a physiologic measure from two groups (i.e., males and females).

Say I obtain a significant difference from an unpaired t-test in blood pressure between men and women. How can I find the percent difference from here? If women have a lower bp, I would like to say something to the effect of "women had 20% lower BP compaired to men P<.05"

  • $\begingroup$ If men have a mean of 180 and women have a mean of 150, then men have, on average, 180/150=1.2 that is 20 % higher BP. $\endgroup$ Sep 6, 2019 at 7:35

1 Answer 1


Implicitly when you do the t-test, you assume that the observations of the two groups are generated from two distributions. When you say a random variable X > 0 with p < 0.05, you're saying that the mean of X is > 0 with probability 0.05.

Similarly, what you're actually trying to say is that the mean of the women's BP is 20% lower than the mean of men's BP. Say the sample mean of men's BP is 100, then this is equivalent to testing if BP from the women's group is equal to 80.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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