Feedback on specific problem from "Understanding Society"?

#1
Wave 5 of "Understanding Society" (the UK general/household panel dataset) has just been released. I've been full of anticipation... But I'm now thinking that I'm not going to find something I expected to find.

For Wave 5, the questionnaire was supposed to ask about citizenship (i.e., whether the respondent holds UK citizenship) again for anyone who previously indicated that they did not hold UK citizenship. (Prior to Wave 5, citizenship was seen as a "stable characteristic", thus asked only in Wave 1 or on entry to the panel; several years ago I persuaded the design team to change this approach, noting that non-citizens can become citizens.)

So, I expected to be able to merge Wave 1 and Wave 5 data and then identify those respondents who had become citizens. As best I can tell, however, it appears that they did not change the survey design in the way I understood they would.

After merging the waves and dropping non-matches, there are now 1438 individuals who were non-citizens in Wave 1. When I tab the citizenship status of these individuals at Wave 5, most (1347) come up as "inapplicable" and the rest (91) come up as "proxy". In my understanding, "inapplicable" means that the question (in this case, citizenship status at Wave 5) was not asked.

(If I tab Wave 5 citizenship without the restriction to Wave 1 UK non-citizenship, I get only 16 substantive answers -- and virtually all of these individuals had "proxy interview" values for Wave 1 citizenship, a form of nonresponse when someone else answered part of the questionnaire on their behalf. So, these respondents are in effect responding to the citizenship question for the first time.)

But before I go back to the design team, I want to make sure I'm not missing something or making a stupid error of some sort. I'd be *very* grateful for any feedback on this matter.

Here is the (very simple) Stata code I'm using for this purpose:

use pidp a_citzn1 using a_indresp.dta
save a_temp, replace
clear

use pidp e_citzn1 using e_indresp.dta
merge 1:1 pidp using a_temp
keep if _merge==3

*to find naturalizers:
tab e_citzn1 if a_citzn1==0

*to see citizenship status at Wave 1 of those who report UK citizenship at Wave 5
tab a_citzn1 if e_citzn1==1

*to see Wave 1 status by Wave 5 answers
tab e_citzn1 a_citzn1