In article <AY0X9.4258$[EMAIL PROTECTED]>, Don <[EMAIL PROTECTED]> wrote:
>I'm currently just concerned with whether the pre-treatment scores of those >in control group, and those in the treatment group are ths same. So for this >exercise, isn't it true to say that I have data for the entire population of >interest, and therefore sigma is known? The scores are either the same or they're not. Most likely not, if the scores are approximately continuous. There's no need for any test. I doubt this is your real question, however. Are you worried that a supposedly random assignment to treatment and control groups wasn't really random? If so, you could do a large number of actually random assignments and compare the difference in scores for the actual assignment with the distribution for random assignments. Or are you worried that even though the assignment was random, it by chance produced unbalanced groups? If so, seeing whether some test or other rejects a hypothesis that you know to be true is not a sensible approach. You just need to analyse the data in a way that doesn't rely on this (eg, include the score as an explanatory variable in a regression model). 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/ . =================================================================
