Take the cross-wave file created yesterday, which contains eleven observations of wJBSTAT and eleven presence indicators (W1 to W11). Turn this into long format with the following list of variables: respondent ID, wave number, presence, job-stat.
In recent versions of SPSS this is relatively easy:
get file 'm:\multi.sav'. VarstoCases /make jbstat from ajbstat to kjbstat /make qfedhi from aqfedhi to kqfedhi /make w from w1 to w11 /index = wave. save out = 'm:\multilong.sav'.
If you look at m:\multilong.sav
you will see, among other
things, many records with no useful information (for waves where
the respondent was absent). These can be dropped harmlessly from
this format, but not from the wide format (you just have to be
careful that you check the wave index number when comparing across
records).