The issue goes somewhat deeper than that. -ksmirnov- computes
incorrect results in the case of ties, and this should be fixed.
However, not that, even if computed correctly, the Kolmogorov-Smirnov
test is conservative for discrete data (see, e.g., Conover 1972).
ben
Conover, W. J. (1972). A Kolmogorov Goodness-of-Fit Test for
Discontinuous Distributions. Journal of the American Statistical
Association 67(339): 591-596.
On 5/15/07, David Airey <[email protected]> wrote:
I have found that this test assumes random sampling. I had someone
attempting it with multiple measures within animals.
While I'm on this topic, I found a note in the Statalist saying that
the implementation of ksmirnov is faulty for discrete data. As far as
I can tell, it has not been fixed, despite Stas and Ben asking:
[...]
Can't we get this one fixed or what??
-Dave
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
