API request to retrieve the Paradata for a single interview


Add an API request to retrieve the Paradata for a single interview.


Paradata contain exclusive information about an interview which under certain circumstances is extremely important. (Example: has the answer to a sensitive question like nationality been deleted/changed?).

Currently Paradata can be exported through the API, resulting in huge files taking up lots of server space and taking a long time to generate and download.
If I need Paradata only for a specific interview I can limit the export request to a time interval just containing a point in time from this interview. But in a census with 10,000+ interviews/day the request will still return many hundreds or even thousands of interviews, resulting in a large overhead (time and space).

Introducing an API request for returning the paradata of a single interview would eliminate this overhead.

Use Case

This is a real world example. For the population census in a multi-ethnic country A, whose territory is also claimed by another country B, there is a strong preoccupation (by the NSO and the World Bank) that for interviewed citizens of ethnicity and language of country B this information will be correctly recorded in the interview, but later changed to A.
Such irregularities have to be detected immediately in order to take appropriate corrective action.

Affected subsystems

Only the new API endpoint, or a modification of the existing paradata export request.


HQ users monitoring an ongoing survey / API implementers.


Population Census


Export a potentially large set of (irrelevant) interview paradata.

What is the effect of this feature?

Reducing the cost (both time and server space) of downloading paradata.


No effect on users who don’t need this API request.


Applicable for situations where only the paradata of specific (suspicious) interviews need to be analyzed.


Doesn’t affect any limits.


There exists another feature request for the same functionality, although for a different use case:

1 Like

How do you identify specific interview? are you able to ‘guess’ what interview(s) must be checked or will end up needing to loop over list of interviews and requesting/checking paradata content for each?

I have to get a list of completed interviews, analyze several aspects for each one,
and identify (‘guess’) a couple of suspicious ones.
Only for those I want to fetch the paradata.

there is an endpoint returning interview history at api/v1/interviews/id/history. Currently it is a bit of a hybrid of outputting a bit more than what users see on the interview history screen from the GUI and less than what is in the exported paradata. Perhaps this endpoint can be extended to cover the use case, maybe with a parameter of returning ‘all’ or only ‘some’ items to serve the both needs of more or less-detailed output.

Hello @klaus , if this is a real World Bank project, please have the TTL to contact me to supply some more details. Thank you, Sergiy

@zurab, I had never used this endpoint. It looks like it does return all (?) the paradata lines for the interview.
If it really does, this request is exactly what I was looking for.

Do you know if it returns all paradata lines, i.e. in the case that an answer was changed, all the answers, not only the last one?

Unfortunately not all and not always - or at least this was my impression when the last time I used it - I was trying to see whether with this endpoint I would replicate the paradata export and I remember reaching the conclusion that I couldn’t.

What I see so far looks good. I’ll verify it for a small survey I have access to.
I only need AnswerSet and status changes.
When I have results I’ll post them here.

1 Like