In a strong sense, neither. -qnorm- is the best "test" of normality,
in my view.
Also, the assumptions of these methods are at most that conditional
distributions are normally distributed, not marginal distributions.
Nick
On Mon, Dec 3, 2012 at 6:16 PM, MIGUEL ANGEL GRUPO DE INVESTIGACIO
ODONT SIMANCAS PALLARES <msimancasp@unicartagena.edu.co> wrote:
> Im trying to conduct several analysis including t test and ANOVA. When i
> try to perform normality analysis using shapiro wilk test it say that the
> data is not normally distributed. But, when i perform shapiro francia test
> (with the same data) it says that the are normally distributed...
>
> Which of these results should I use (Wilk or Francia)?
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/faqs/resources/statalist-faq/
* http://www.ats.ucla.edu/stat/stata/
