Make EthicsApproval non-optional

[yarikoptic]

Per stand-up discussion.

Might need fixing etc

[bendichter]

I see that it is important for us to ensure proper ethics approval. I just want to make sure we do it in the a good way, and I am not sure that only requiring an EthicsApproval number is the best way.

Let's consider a re-user coming across a dandiset with some approval number: "234jnjo2234k" and having some questions as to the legitimacy of the approval process for this dataset. Without any information about the approving body, there would be no way to verify, investigate, our ask questions about this approval. Maybe they could infer this body by looking up the institution of the lead author, but institutional affiliation is not a requirement, and even when it is present this would not always lead to the correct approving body. Papers I see online say something like "All procedures were approved under NYU School of Medicine IACUC protocols." I find this more useful, because it would allow a user to contact that organization to confirm that the approval is legitimate and ask any questions they may have.

My suggestion would be either to require approving body (e.g. "NYU School of Medicine IACUC") or to require both approving body and protocol number. The protocol number on its own is in my opinion insufficient.

[djarecka]

just an idea to @bendichter comments. In datacite that would be under Rights I believe, see p.27 The main field is "Free text, Provide a rights management statement for the resource or reference a service providing such information" and in addition you can have rightsURI, rightsIdentifier, rightsIdentifierScheme, schemeURI

[djarecka]

@bendichter - actually, in our schema we have already contctPoint, so this is probably what you have in mind

[bendichter]

@djarecka yes, but the contact point does not have the name of the organization and appears to be optional.

[djarecka]

yes, this is optional now, but probably we can make this required and the identifier optional (to align also with @TheChymera point of view)?

The description says Information about the ethics approval committee., but we could perhaps be more specific

[satra]

My suggestion would be either to require approving body (e.g. "NYU School of Medicine IACUC") or to require both approving body and protocol number. The protocol number on its own is in my opinion insufficient.

how about this? we change it to require approving body and an optional protocol number.

[TheChymera]

@satra that's pretty good I'd say. Approving body would be a string, and nothing we need to validate or index, right?

[satra]

ping @yarikoptic - do you want to pull this out of draft?

[satra]

i'm going to close this as i want to separate the dandi requirements from type validation after pydantic 2.0 is merged.