Re: Normality Test


[ Follow Ups ] [ Post Followup ] [ Netlib Discussion Forum ] [ FAQ ]

Posted by Massimo Gaspari on April 16, 1998 at 05:17:09:

In Reply to: Normality Test posted by Steve McCormack on April 08, 1998 at 13:37:09:

: Hello,

: I am not a mathematician but need help in applied statistics.

: Does anyone know of a simple statistical test (or combination of tests) that will score how closely a set of data matches the frequency distribution of the normal distribution. ? I'm aware of the histogram but that only gives a visual cue as to its shape, not a discrete numeric scoring of the entire dataset, plus the shape you get is driven by the bin size you choose.

: The test I'm after would return a perfect score of say 1.0 when given data that exactly matches the frequency distribution of a normal distribution. Conversely, it would also return a zero (the worst score possible) when given data that is as anti-gaussian as you can get.

: Any help would be appreciated.

: Steve


Steve,

Look any statistical textbook. Chi-square test
may help you.

If you need help to apply it feel free to e-mail
me.

Bye



Follow Ups: