You are viewing a single comment's thread from:
RE: How Use Excel T.TEST - Was a Marketing Campaign Successful?
T-Tests are for data we can assume is normally distributed. That applies for much human data, but isn't always true.
T-Tests are also not useful for ordinal (ranked, but the number doesn't represent the distance between the ranks) or nominal (category) based data.
The most common misuse I see of T-Tests is when examining Likert-like data (e.g. 0 = Never, 1 = Rarely, 2 = Sometimes, 3 = Often, 4 = Frequently, 5 = Always). This data is ordinal because it is ranked but the distance between ranks is not clear. For this kind, of data use something like: Wilcoxon or Mann-Whitney.
T-Test is still a great test to use!
Hay and thanks for visiting. Yes this is true, T.TESTS are not good for ordinal data at all. However it is good for the examples shown. I must check to see if excel has built in functions for Wilcoxon, although it might be a combination of functions.