Re: test of normality



thierry wrote:
> >I tested a normality (thanks for all you) and I have :
> >coef of kolmogorov-Smirnov = 0.293 (sig = 0.000)
> >skewness = 10 (std. error = 0.012)
> >kurtosis = 326 (std. error = 0.025).
> >what should I do ?
>
> Are you sure about the skewness and kurtosis? The standard
> errors may be reasonable, but those figures for skewness and
> kurtosis are EXTREMELY high. It is hard to get sample skewness
> and kurtosis that large.
>
> Excuse me, I have these values (on SPSS : analyse -->descreptive
> analysis --> explore):
> coef of kolmogorov-Smirnov = 0.293 (sig. = 0.000)
> skewness = 8.569 (std. error = 0.012)
> kurtosis = 134.309 (std. error = 0.025)
> but it is still high . What do you thonk about ?
>
> Thierry, habibidmd@xxxxxxxx

I thonk you should not use ANY of those measures for your purposes
of checking normality (or lack thereof).

-- Bob.

.

Relevant Pages

  • Re: significance of skewness
    ... The true underlying distribution of this data is unknown to me. ... >> I want to determine whether the skewness of the data is significant. ... > There is a literature on "robust tests of symmetry". ... robust measures of skewness and kurtosis. ...
    (sci.stat.math)
  • Kurtosis, Skewness and the Cummulant Generating Function
    ... formulas for skewness and kurtosis that I am familiar with are ... skewness and kurtosis of a continuous distribution, ... information on the cummulant generating function, ...
    (sci.stat.math)
  • Re: Different skewness while keeping kurtosis constant
    ... It is possible to change the kurtosis independently from the mean, ... and skewness using the Pearson VII family of distributions. ...
    (sci.stat.math)
  • Re: Multiple Regression doubts
    ... Maybe normalizing is a wrong word, ... Any linear transformation will preserve skewness and ... kurtosis, so that's no problem. ...
    (sci.stat.edu)
  • Re: random number generation
    ... >>> chosen skewness or kurtosis. ... then you use that distribution for the ... >> Kurtosis in advance, then match the skewness. ... >> John D'Errico ...
    (comp.soft-sys.matlab)