Merging PISA waves in Stata

#1
I am running into trouble trying to merge multiple PISA waves into one Stata file for running some hierarchical models. So far, I have combined the student and school files, but do not know the best way for merging 2000 through 2012. The problem is further complicated by the fact that variables in one year do not correspond to the same measure in another year.

Should I be merging and then reshaping, or looking at the append and joinby commands?

Thanks in advance for your response!!
 

maartenbuis

TS Contributor
#2
the rows in a dataset represent the same observation of the same person. PISA doesn't repeatedly ask the same students, so you don't want to put observations from different waves on the same row. Putting observations on the same row is what merge is for. So that is the right command for linking information from students and schools, but the wrong command for appending waves. Instead, you should look at the append command. That way you add observations from the different waves, i.e. the waves appear underneath each other, which is exactly what you want.
 

Lazar

Phineas Packard
#3
Not only do they not ask the same student from cohort to cohort but the "major domain" changes from cohort to cohort (OECD seem to change a fair bit of other stuff other than the major domain from cohort to cohort as well). So for many (most) variables not even append makes a lot of sense.