>doing all this inferential work is an unnecessary step in what appears to >me to be a rather simple problem ... IF the criterion set by EEOC is 2 sds >different ... (their criterion is not two standard ERRORS different)
It seems highly unlikely to me that the EEOC criterion is a difference of twice the standard deviation of the salaries. That would usually be an enormous difference in means (in this context). They surely must be interested in how the difference compares to the standard errors. This discussion is absurd. It's not possible to rationally defend a position of ignoring statistical significance in this context, unless your aim is to simply declare discrimination to exist regardless of whether there's any reason to believe it does. The example I gave of only two employees (male earning $70,000, female earning $65,000) makes quite clear the need for considering statistical significance. Situations with more employees differ only in that the need for assessing statistical significance is perhaps less obvious, though no less real. Radford Neal ---------------------------------------------------------------------------- Radford M. Neal [EMAIL PROTECTED] Dept. of Statistics and Dept. of Computer Science [EMAIL PROTECTED] University of Toronto http://www.cs.utoronto.ca/~radford ---------------------------------------------------------------------------- . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
