If I am comparing the effect of age, sex, location (numerical) on various numerical parameters, but one of these responses are not normally distributed, can I still do a factorial ANOVA?

I can't do a Kruskal Wallis test as I am comparing so many variables.

EDIT: Anderson-Darling normality test gave a p-value of <0.005

  • $\begingroup$ What exactly isn’t normal? $\endgroup$
    – Dave
    Jul 1, 2020 at 14:46
  • $\begingroup$ One of my responses isn't distributed normally. $\endgroup$
    – Sam
    Jul 1, 2020 at 14:50


Browse other questions tagged or ask your own question.