Peter L. Flom, PhD Assistant Director, Statistics and Data Analysis Core Center for Drug Use and HIV Research National Development and Research Institutes 71 W. 23rd St http://cduhr.ndri.org www.peterflom.com New York, NY 10010 (212) 845-4485 (voice) (917) 438-0894 (fax)

>>> "Swank, Paul R" <Paul.R.Swank@UTH.TMC.EDU> 10/14/05 4:00 PM >>> wrote <<< I only made the statement because of the number of people who are trashing significance tests because, they say, people misstate their meaning. No offense intended but those folks are not magnanimous. >>>

with the statement being <<< I'm sorry but "if the KS statistic exceeds 1.63 there is only a 1% chance that the distributions are really the same" is not accurate. If there were no differences in the distributions then we would expect a KS statistic of 1.63 or greater only 1% of the time. >>>

This statement is entirely correct. But what about people who trash significance tests for other reasons? I, myself, am fairly far on the 'anti-significance test' side.....mostly because I don't see many cases where they answer the question we want to ask. i.e., they ask (as Paul notes) "how likely are these data, given the null?" whereas what we usually want to know is "How likely is the null, given these data?"

Of course, I still use p values, because journal editors and Principal Investigators like them.

The whole sig. test issue has become a big controversy in psychology

Peter