I have measured estrogen concentration in blood of 7 animals over 8 time intervals. Before I run repeated measures ANOVA, should I test for normality within each time, or across (pooled) overtime i.e. 8 checks one per time, or one check for the data as a whole. Thanks.

  • 1
    $\begingroup$ You have a very small sample. Testing normality would be difficult and possibly not necessary. The type of data you have is called longitudinal or panel data. Check that literature. You may be able to take a nonparametric approach. $\endgroup$ Commented Apr 11, 2017 at 0:29


Browse other questions tagged or ask your own question.