Gus Gassmann said on 9/30/02 7:35 AM: > >Evidently I am either ignorant of what I am doing or using a different >definition >of causality. I ask again: Does causation transfer to a subset? And please >try >to answer this question without reference to CR. That has nothing to do >with it!
Maybe some slight rewording will help both of you get to the same page. Maybe the question should be, "Does causation transfer to a representative subest?" Representative needs to be defined, but is probably best defined as a simple random sample of the original data used to investigate causation (which was, presumably, collected using rational methods). If a simple random sample of the original data to create a subset is an insufficient definition, Bill, please offer what it needs to be. A simple random sample to create the subset would typically have the same distribution as the original data which you've already used to investigate causation or lack thereof, therefore already has characteristics you favor. Paul . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
