Re: Determining if samples are normal

[Mike Miller]

On Tue, 27 Sep 2005, Henry F. Mollet wrote:

> Going back to the original question and quoting from a manual: A 
> histogram of the data (corresponding to the pdf) is a rather poor method 
> of determining whether data look normally distributed. More powerful is 
> a plot of the values of the variable against corresponding percentages 
> of a standard normal variable (Gnanadesikan 1977). If the data is 
> normally distributed, one should get a straight line.

That's what we were doing in the q-q correlation test.  We didn't look at 
the graph, but we computed the correlation coefficient to see if it was 
smaller than one would expect by chance for genuine normal data.


> Then, I guess even more powerful, are all the test mentioned, including 
> anderson_darling_test

For triangular data, the q-q correlation test was more powerful than 
Anderson-Darling in a comparison that Paul Kienzle and I did.

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
-------------------------------------------------------------