Understanding Society User Support

Hello,

I would like to use pidp for the household reference person in data file c_indall, but it seems that the variable c_hrpid does not in general match pidp for any member of the household. I can work around this by using c_hrpno instead, but I just wanted to check whether there might be an error with c_hrpid or whether I have misunderstood something.

I used the following Stata code:

rename c_* *
gen x=hrpid-pidp if pno==hrpno
tab x
sort hidp pno
brow hidp pidp hrpid pno hrpno x

I expected x=0, but this is true for only 9 cases. Here are the data for the first 4 households.

hidp pidp hrpid pno hrpno x
68013604 68004087 68004088 1 1 1
68020404 68006127 68006128 1 1 1
68020404 68006139 68006128 2 1
68020404 68006135 68006128 3 1
68020404 68020564 68006128 4 1
68040804 68006807 68006808 1 1 1
68047604 68007487 68007488 1 1 1
68047604 68007491 68007488 2 1
68047604 68007499 68007488 3 1

I have not checked all w_indall files, but found the same problem for wave 'e'.

Thanks very much.

Fiona