How to report occurrences derived by interview?

[CecSve]

The GBIF Secretariat has received a question about how to share occurrence records derived from an interview:

I am wondering how to report occurrences that were derived by interview.

For example, when conducting a survey the data collector will record direct observations of occurrences but often times will also report observations of occurrences made by other individuals that they interviewed.

I would like to be able to distinguish between direct observations and indirect (interview based) observations to be able to give data users that information.

Any thoughts on how to do this?

I wonder if someone in the DwC community can offer advice on this? I remember there were some discussions around this for GBIF TG for indigenous data governance.

[tucotuco]

Interesting. In terms of Simple Darwin Core (and therefore Darwin Core archives) the dwc:basisOfRecord would still be HumanObservation in the absence of a more specific value for cases like this. The person who made the original observation (not the one who recorded it, if they are different) should be the one in dwc:recordedBy according to that term's definition.

What's missing with the above is that there is no indication that the information came from an interview and the identity of the interviewer is lost. What to do about this depends on what is trying to be achieved. If the goal is simply to not lose that information, dwc:occurrenceRemarks can cover that. The dwc:recordedBy term could also contain the interviewee first in a list with the interviewer second. If the goal is to have that information be a little more structured, then dwc:dynamicProperties could be used with keys to indicate something like sourceType "interview" and interviewer with the person's name. If the goal is to distinguish interview-based observations from other human observation types (such as literature-based), then proposals should probably be made for additional recommendations to the vocabulary for dwc:basisOfRecord, providing the justifications required by the Vocabulary Maintenance Specification.

In Darwin Core Data Package the distinct roles of the interviewer and interviewee could be shared as distinct records of OccurrenceAgentRole.