Suppose we have two samples of a variable taken under different conditions: e.g. A1 without medical treatment and A2 after medical treatment. These are not necessarily normally distributed. Suppose A1 ranges from 0 to 10 and A2 ranges -2 to 15. Now how do I test for the significance of these range increases?
Any thoughts? Thanks in advance.
$\begingroup$
$\endgroup$
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
My suggestion would be...
Sample range distribution can be calculated. when X1 ~ Sample Range distribution from N(mu1, sigma^2), X2 ~ Sample Range distribution from N(mu2, sigma^2) we can get X1/X2 distribution much like F-distribtuion.
Then we can test the significance of the range increase(range1/range2) with that distribution.
