On 17 Aug 2003 at 13:37, wuzzy wrote: Wouldn't it be possible to try some sort of (multiple) imputation guided by the correlation matrix?
Kjetil Halvorsen > This question is too basic to ask in the sem groups kindof off topic > here so I may not have much luck but: > but how do you set up your data if *part* of it is based on a > correlation matrix? > > I have original data on all except one variable and would like > to know how my model works with this variable thus would like to work > with a correlation matrix. > > I cannot think of how to set up the data in a correlation matrix form > for Amos to interpret (eg., which would be the rows and columns?) > Thx.. > . > . > ================================================================= > Instructions for joining and leaving this list, remarks about the > problem of INAPPROPRIATE MESSAGES, and archives are available at: > . http://jse.stat.ncsu.edu/ . > ================================================================= . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
