0

I have a control group defined by a certain variable (say values 0-5) and an sick group (values between say 5-20). Within each group the data are normally distributed, but (off course) not when looking at the the whole cohort. Am I allowed to use parametric tests in this instance? Sorry for both my English and my lack of statistical skills...

Sara
  • 1
  • 2
    Welcome to Cross Validates! It’s a little bit hard to tell what you’re doing, but normality assumptions have to do with the in-group distributions, not all of the data pooled together. It sounds like you are okay to assume normality. // As a heads up, “parametric” testing does not require normality. It is totally fine to do a parametric test for other distributions, just with a different parametric model like Poisson or Binomial. – Dave Aug 16 '21 at 22:42

0 Answers0