Re: Determining if samples are normal

[Mike Miller]

On Mon, 26 Sep 2005, Przemek Klosowski wrote:

>   You warp it by the normal curve so that 'normal' data will produce a
>   streight line.
>
> That's very clever, but it's not a very sensitive test: for 
> y=rand(100,1)+rand(100,1) nt(y) returns 0.996 (this holds even when one 
> uses 10000 points for better stats). The distribution of rand+rand is 
> triangular, so noone would mistake it for a gaussian.


For a sample of only 100?  The best way of deciding would be to compare 
methods.  I like the idea of using rand(100,1)+rand(100,1) because it is 
unimodal, symmetric, easy to produce and not normal.

Mike



-------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.

Octave's home on the web:  http://www.octave.org
How to fund new projects:  http://www.octave.org/funding.html
Subscription information:  http://www.octave.org/archive.html
-------------------------------------------------------------