The simulated data set for the EMS example (see the link under 'EMS group exercise' for more details) was created under the assumption that 3 hospitals, 3 methods, 4 doctors (for each hospital by method combination) and 2 patients (for each doctor) were present in the study. If C1 contains the hospitals, C2 contains the method, C3 contains the doctors, C4 contains the patients and C5 contains the simulated response, we generate the data as follows: set c1 (1:3)24 set c2 3(1:3)8 set c3 9(1:4)2 set c4 36(1:2) rand 72 c5 Before adding column names, you can save this as a tab-delimited text file (or choose the delimiter of your choice) to be read into SAS. To set up the file for reading by the commands under the link 'SAS code' for the EMS example, you select 'Save Worksheet' under the 'File' menu and choose 'Text' as worksheet 'Type'. Under options, you'll want to select 'None' for variable names and 'Tab' as the single character delimiter. Technically, the SAS program expects the file to be named 'a:\stat706\ems.sas', but you may have to adjust that accordingly. Within Minitab, you should add names to the columns. Under 'Balanced ANOVA', enter the response column and model statement. The model statement should be: c1|c2 c3(c1 c2) Include c1 and c3 as random effects. Under 'Options...' ask to display the EMS table and select the restricted form of the model (you'll obtain the incorrect F test for Method otherwise). If you try to analyze the model under Generalized Linear Model rather than Balanced ANOVA, the restricted form of the model cannot be selected as an option and the EMS will not be computed correctly.