mismatch merging indall m:1 using hhresp (merge=1)
I was merging household-level variables onto individual files (indresp) and found mismatches, so to make sure I had not made any mistake, I tried again with _indall (as I understand all individuals there do have a household). I'm not sure is
As an example, for wave 2 I found 14 individuals with no information at the household level. Some of them were OMS and had been full-interviewed (ivfio=1). See syntax below.
// open indall and merge household level data for all individuals
use pidp b_hidp b_pno b_sampst b_ivfio using "$ukhls/ukhls_w2/b_indall", clear
// merge household level variables onto individual level file
merge m:1 b_hidp using "$ukhls/ukhls_w2/b_hhresp"
tab b_sampst if _merge==1
tab b_ivfio if _merge==1
I checked the hhresp file and looked for some of those b_hidp numbers and did not find them there.
I am doing this for all samples and the number of those individuals with merge=1 exceeds 2,000 for wave 10.
What am I doing wrong? or why some individuals have no hh information? Particularly when they are OMS & ivfio=1.
I might have misunderstood the file contents.
Updated by Alita Nandi over 1 year ago
- Status changed from New to Feedback
- Assignee changed from Alita Nandi to Pilar Zueras
- % Done changed from 0 to 90
- Private changed from Yes to No
While all individuals in indall must be in a responding household, the households in the file hhresp are only those households which have completed a household questionnaire. There are some cases where the household grid was completed (so in indall) but a household questionnaire was not completed (so not in hhresp).
Understanding Society User Support Team