I have hundreds of variables each has more than 100000 observations. My aim is to determine which variables ara normally distributed. Usual tests like Shapiro-Wilk tests are likely to reject normality of distribution even the distribution is really normal. Since I have hundreds of variables, I can't use graphs to decide normality. How can I do such a mulpitle test using R? I will be very glad for any help. Thanks a lot.
Asked
Active
Viewed 20 times
0
-
See this answer http://stats.stackexchange.com/questions/92374/testing-large-dataset-for-normality-how-and-is-it-reliable?rq=1 – mdewey Apr 11 '16 at 17:35
-
Translated into simple terms, this post appears to ask, "I would like to find normally distributed variables. I have a procedure that does a very good job of that. What are my alternatives?" I hope that makes it evident that something important is missing from the question: namely, *why* are you doing this and what specifically is the matter with being able to do it well? – whuber Apr 11 '16 at 17:40