On 20 Feb 2003 11:48:08 -0800, [EMAIL PROTECTED] (saisat) wrote: [snip. About multiple groups: different means, SDs, Ns] > > What is the most accurate way to determine the overall mean and the > standard deviation of the above disributions. > > Also, I need to remove any distributions that are not similar to the > rest of the distributions in the "peer" group. > The first sentence above would be pretty simple, if it were not for the subsequent sentence that makes it impossible for anyone to give you good professional advice.
Hardly anybody throws away some groups in order to look at the means and SDs of the groups that they 'like'. The people who do something similar, will fall into one of two groups, about 95% of the time -- 1) Folks who understand the meaning and implications of dropping groups; almost always, they will have no trouble applying the ANOVA formulas, back-solving to get the practical solutions; so they would never ask this question. 2) Folks who are making a big mistake, from ignorance. So, in my estimation, there is only a small chance that you need those answers without having a more serious need for consultation -- on models, etc. If you do: Look for the formula for one-way ANOVA, unequal Ns, in any basic textbook. It should not be too difficult to work backwards to get the pooled SD, especially after you recognize that it is the Total Sum of squares divided by the Total DF. -- Rich Ulrich, [EMAIL PROTECTED] http://www.pitt.edu/~wpilib/index.html . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
