[Nick Cox]

I have no advice here, except that I can't imagine a reason to privilege the minimum rather than the maximum.

Inconsistencies could mean (1) a change of ethnicity, which may be allowed or happen, depending (2) a mistake of some kind. I have no idea how you distinguish. Changes of reported ethnicity may reflect changes of perceived identity either independently or after marriage, etc. I can't advise on what makes sense in your country. The question can be very sensitive, raising issues of stigma, status, religious or political implications, etc.

[Shamsudini Amidu]

Hi Nick Cox and Team,

Still waiting for your comment.
What if the min is not equal to the max? what will be the best approach to use in filling the missing values in ethnicity2 because I can still see a missing value in ethnicity2 for those whose min is not equal to max.


I still appreciate your time.

Thank you.

[Shamsudini Amidu]

Hi @Nick COX,
Sorry just saw your comment!

[Shamsudini Amidu]

Hi @Nick COX,
I appreciate your response.
Well for all that I know, here in Ghana, Ethnicity is just similar to Gender where if a respondent reported that he is a male in wave1, we don't expect to see different Gender categories appearing in either of the waves(wave2 or wave3) is should be the same.
So after making a follow up to the reason's why there is that inconsistency, I established three(3) reasons, :.
1. Enumerator made a mistake in selecting the wrong response.
2. Different respondent respond on behalf of another household member who was absent in the next round.
3. Enumerator skip the ethnicity question without any reason(causing the missing gap)

Is there any approach I can take assuming we don't allow changes to ethnicity sir ?

Thank you.

[Nick Cox]

If you don't allow changes in ethnicity, then I don't have any suggestion except to ignore people for whom the data are contradictory.

I don't have any specialist expertise here. You need to talk to social (medical?) scientists in Ghana to get more advice, not least any advisors or supervisors you have.

[Shamsudini Amidu]

Hi @Nick Cox,
I am most grateful for your comment.
It's helpful.
Noted.
Thanks a lot.