Can I use the t-test to compare means between groups (N=300) if data is not normally distributed?
Asked
Active
Viewed 123 times
0
-
1It depends. In some cases, n=3 suffices even though the distribution is non-normal. In other cases n=3000 isn't enough... – Glen_b May 15 '13 at 09:43
-
2See any of these (only a sampling of many): 1 2 3 4 5 – Glen_b May 15 '13 at 09:54
-
Thanks a lot, Glen. This is really helpful. Did not see the other posts before ... – user25630 May 15 '13 at 10:33