Why, for goodness' sake? What did you think you'd do with the information if you found it? (Further comments embedded below.)
On Fri, 20 Feb 2004, Tanguma, Jesus wrote (edited): > I am trying to locate sources on Type I, II, III and IV (if any exist) > error rates. > > I do have some sources on Types I and II, only two on Type III and > none on Type IV. Does this surprise you? Type I and Type II errors are at least defined so as to permit of a sensible definition of "error rate", in the form of conditional probabilities that depend on the hypothesis/es being tested and on the distribution(s) applicable to the test(s) being made. As rejoinders to your earlier post should have made clear, Types III and IV are little more than plays on words, depending almost wholly on the investigator's intention(s). About the only way one could get an "error rate" for either is by counting the number of instances observed and dividing by the number of occasions when it might have been observed (which would presumably be the size of the whole corpus of relevant research or analyses); neither of which numbers is well enough defined (especially the rather metaphysical number that depends on the inquirer's intentions) to admit of a sensible measure of "error rate". -- DFB. > Jes�s Tanguma > University of Texas-Pan America > CIS/QUMT Dept. > MAGC 3.328 > 1201 W. University Drive > Edinburg, TX 78541-2999 > [EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]> ------------------------------------------------------------ Donald F. Burrill [EMAIL PROTECTED] 56 Sebbins Pond Drive, Bedford, NH 03110 (603) 626-0816 . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
