I have a question regarding Fisher's > > > exact: is it a legitimate test to run on a particular set of data I > > > have. > > > No statistical test is legitimate in the hands of the inexperienced.
No statistical test is legimate in the hands of the experienced when they have an agenda. Mark Twain said there are three kinds of lies - "Lies, damned lies and statistics". It is also why the bottom of my business card used to read "Numerical Advocate" rather than statistician. Lest anybody challenge this as off-topic, there is a wonderful package - 'R' and a somewhat harder to find "pspp" both of which have endless tests to a amuse and befuddle and both of which run on mandrake. My advice, run every test you can find, smile non-chalantly, and pass the burden of legitimacy on to the consumer. Jim Tarvid PS. I really enjoyed the non US contributions to the thread on nature and character of President Bush and his "fatal attraction" with Saddam Hussein.
Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com