Hi. I made up an example that you should be able to generalize from. Run it and see if that is the kind of plot you are envisioning (of course, yours would have more lines):
data list free /beforeA afterA beforeB afterB beforeC afterC.
begin data.
12 09 05 07 14 10
18 12 09 10 16 11
10 06 10 13 20 10
13 10 08 09 18 09
17 13 12 15 17 12
end data.
GLM beforeA beforeB beforeC afterA afterB afterC
/WSFACTOR=time 2 Polynomial ABC 3 Polynomial
/METHOD=SSTYPE(3)
/PLOT=PROFILE(time*ABC) TYPE=LINE ERRORBAR=NO MEANREFERENCE=NO YAXIS=AUTO
/CRITERIA=ALPHA(.05)
/WSDESIGN=time ABC time*ABC.
You have two times (before, after), so time=2. Your variables would be in two 12-var groups "all 12 before", "all 12 after", something like BEFORE_01 to BEFORE_12 AFTER_01 to AFTER_12.
Where I have the generic "ABC" you would have some sensible name that identifies what you're measuring.
------------------------------
Rick Marcantonio
Quality Assurance
IBM
------------------------------