Deleting cases

[irvin]

I am trying to predict the presence/absence of future depression from a measure of past working conditions. I would like to omit from the analyses individuals with prior depression. I used the Omit selection in the Select or omit Cases item on the Data Manipulation menu to delete the dependent variable (later depression when the worker was depressed earlier) and two other variables and place them in outlying columns. I executed the procedure and received no error messages.

I don't know why this procedure did not work. I am new to the procedure, usually organizing my data in SPSS before writing the data and having MLwiN read the data. But I don't want to go back to SPSS every time I hit a stumbling block, and have learned some MLwiN syntax. But I am hitting a wall in omitting data. I would appreciate getting some help.

Thanks.

Irvin

[ChrisCharlton]

MLwiN handles data differently to SPSS in that rather than assuming that the data is rectangular with a row per case and a column per variable it instead allows columns of different length and only requires that columns in the model have the same length. It is therefore possible that you only removed records in the three columns that you selected rather than in all the variables in the model.

An alternative to removing cases that you want to omit is to generate a zero/one column with a one for cases you wish to exclude and zero for cases you want to use. You can then use this column to exclude cases from the model by either choosing:

Model->Hierarchy Viewer
Click Options
Tick Exclusion on
Choose the generated column
Click Done

or by using the EXCL command, which is described in the help system.