The W1 to W11 variables tell you about the presence of the individuals in all eleven wINDRESP files. Look at the variables to get a sense of the different sorts of participation: present at all waves, present in the beginning but dropping out for ever, entering after the first wave, dropping out for a period and returning.
Use VECTOR/LOOP to summarise this information. First, just count the numbers of interviews each individual gives (there is an easier way with the COUNT command, but use LOOP):
get file "m:\multi.sav". compute count = 0. vector w=w1 to w11. loop #i=1 to 11. . if w(#i)=1 count=count+1. end loop. freq count.
A more complicated example: how many times do people enter the panel after wave 1? Entry will show up here as a Wn variable equal to 1 after all zeros. This is a slightly more complicated loop:
compute enter = 0. vector w=w1 to w11. do if w1=0. . loop #i = 2 to 11. . if w(#i-1)=0 and w(#i)=1 enter=#i. . end loop if enter > 0. end if. freq enter.
Think carefully about the loop: (i) why does a value of ENTER greater than zero not only mark a late entrant, but also give the wave number of entry, and (ii) why does the loop not also wrongly pick up people when they re-enter after a temporary gap? (iii) Why is the fragment ``if w(#i-1)=0 and ...'' redundant?
Now write your own loop to identify when people first drop out (ignore any subsequent re-entry). A drop-out can be identified as the last wave (w(#i-1)) being 1 and this wave (w(#i)) being 0.
© Brendan T. Halpin (e-mail), GNU Free Documentation Licence