testing two means vs two medians. I often deal with data that isn't normal. In &quote;Six Sigma and Minitab&quote; you speak about the 2 Sample t-test can be reliable if the data is &quote;roughly&quote; normal. This is very fru



testing two means vs two medians. I often deal with data that isn't normal. In "e;Six Sigma and Minitab"e; you speak about the 2 Sample t-test can be reliable if the data is "e;roughly"e; normal. This is very frustrating. Your position is wishy washy. Moving to a median test seems less insightful than testing means so I prefer the t-test. What do I do if the t-test rejects Ho but the median test accepts Ho? What do you mean by "e;roughly"e; normal?