Just a small pedantic point, but non-skewness does not imply symmetry. Symmetric distributions have all odd moments (about the mean) equal to zero; not just the skewness. In practice, however, non-skewness and symmetry often go together, at least for unimodal distributions.
As for how small is 'small': well it depends what your aim is I suppose.
Regards,
Chris
________________________________
From: A UK-based worldwide e-mail broadcast system mailing list on behalf of giov
Sent: Wed 27/08/2008 15:35
To: [log in to unmask]
Subject: evaluating skewness values
Hi all,
I computed skewness of sample populations (I have 15 populations) and I
founded 15 values ranging
from -0.05 to 0.88. I am interested in evaluating the symmetry of my
populations (and consequentially the normality of these). I know that if the
skewness value is small, then the distribution is simmetric. But, what means
"small"?
thank you!!
giov
--
View this message in context: http://www.nabble.com/evaluating-skewness-values-tp19182327p19182327.html
Sent from the AllStat mailing list archive at Nabble.com.
|