Improving guidelines for author affiliations

[arfon]

There is some confusion around what an acceptable author affiliation looks like for a JOSS paper. This issues is designed for us to resolve this question so that we can improve our docs.

I'll take a stab at defining what I think our expectations should be here:

Authors at a research institution

• Department, Institution, Country e.g. Department of Chemistry, University of Nottingham, UK

Authors at a business/non-research institution

• Business/organization name, Country e.g. GitHub Inc., USA

Non-affiliated authors

• Affiliation, Country e.g. No Affiliation, USA or Independent Researcher, USA

What do you all think about this @openjournals/joss-editors? In particular cc @kyleniemeyer @Kevin-Mattheus-Moerman who I have seen discussing this.

/ cc https://github.com/openjournals/joss/pull/598

[danielskatz]

Maybe I'm feeling a little argumentative this morning, but I don't really think we need to prescribe or enforce consistency here. I am happy for us to provide examples though.

I also wonder why we care. What's the purpose of an affiliation? If it's to identify the author, doesn't the ORCID do that well enough? Maybe it would be better to use the ORCIDs we have, and add ORCID (id) symbols after author names that link to their ORCID profiles, as other publishers do (e.g. https://f1000research.com/articles/7-1599/v1)

Finally, who are we doing this for? If it's for the authors, we could just let them do what they choose. If it's for us, because our internals require an affiliation, we can just say that and let the authors again fill it in as they choose.

Specific thoughts:

1. Some may want to provide departments; other may not. For me, this could range between

   University of Illinois at Urbana-Champaign, USA

and

National Center for Supercomputing Applications; Department of Computer Science; Department of Electrical and Computer Engineering; School of Information Sciences (iSchool), University of Illinois at Urbana-Champaign, United States

(noting that the iSchool is a school without departments, not the school that contains CS and ECE). Is this really needed for my paper?

And if it is, how do we deal with multiple departments in one institution (semicolons? ampersands?), or multiple institutions. What if I also want to list my Argonne affiliation?

2. I think country abbreviations are fine - I don't see a need to spell out countries.

3. If someone doesn't have an affiliation, why do we need make them add text to say so? And if someone does not have an affiliation, why do we care what county/ies they live in?

[arfon]

Maybe I'm feeling a little argumentative this morning, but I don't really think we need to prescribe or enforce consistency here. I am happy for us to provide examples though.

:smile:

I also wonder why we care. What's the purpose of an affiliation?

One reason I think we care is that it's useful to be able to geocode the string they provide in some kind of automatic way (for our analysis). This to me suggests we need to have a country name or abbreviation (e.g. USA).

Some may want to provide departments; other may not.

Agreed. Perhaps we don't need departments. I don't feel strongly about this.

I think country abbreviations are fine - I don't see a need to spell out countries.

Agreed.

If someone doesn't have an affiliation, why do we need make them add text to say so? And if someone does not have an affiliation, why do we care what county/ies they live in?

Again, for tracking/journal-level metrics.

[danielskatz]

If someone doesn't have an affiliation, why do we need make them add text to say so? And if someone does not have an affiliation, why do we care what county/ies they live in?

Again, for tracking/journal-level metrics.

If this is our goal, then maybe we need to use a formatted means to collect this, rather than plain text? This would also let us validate it. Maybe we should break country from the text affiliations part of the paper.md? So instead of

 - name: Lyman Spitzer, Jr. Fellow, Princeton University
   index: 1
 - name: Institution Name
   index: 2
 - name: Independent Researcher
   index: 3

maybe

 - index: 1
   unit: Lyman Spitzer, Jr. Fellow
   institution: Princeton University
   country: USA
 - index: 2
   unit: Department Name(s)
   name: Institution Name
   country: Country Name
 - index: 3
   name: Independent Researcher
   country: Country Name

where maybe we say that country is required but nothing else is?

We could then validate any parts of this we wanted to as part of the processing. e.g. "'United Staatees' is not a recognized country"

[mikldk]

I get the idea which I in principle support, but I cannot help but think that country without affiliation to me seems a bit strange - what does country then refer to? The person's home address, the workplace's, Citizenship (which?) etc. With affiliation it is a lot easier as it is the country of the affiliation. Here it will probably often be together with an institution so maybe it's not a real issue.

[kyleniemeyer]

I've been interested in having some standards around affiliations for a while, for two reasons:

• a standard format makes it much easier for us to gather data on who we are publishing; unfortunately, ORCID doesn't totally solve this, because fields are not necessarily visible / required by default
• To me there is a professionalism standard that I want JOSS to meet—I would like our papers to look like other academic journals in certain ways. This is one way to continue demonstrating that we are a serious journal

[kyleniemeyer]

I'd be happy to adopt a standard like what @danielskatz sketched out above

[arfon]

/ cc @tarleb for visibility - this could be part of our work to extend the metadata that JOSS (and Pandoc) support.

[danielskatz]

I get the idea which I in principle support, but I cannot help but think that country without affiliation to me seems a bit strange - what does country then refer to? The person's home address, the workplace's, Citizenship (which?) etc. With affiliation it is a lot easier as it is the country of the affiliation. Here it will probably often be together with an institution so maybe it's not a real issue.

I guess this to me would be the country they work in. If they work in multiple places, they could use something like

 - index: 1
   name: Independent Researcher
   country: Country 1 Name
 - index: 2
   name: Independent Researcher
   country: Country 2 Name

[tarleb]

Could we also include city? Not all institutions contain the location in their name. It might make sense to have a separate field for it; otherwise that info would be folded into institution.

Slightly tangential, but how would people feel about using scholarly-metadata.lua to give authors more flexibility when entering metadata?

[danielskatz]

It's not clear to me that we have a use case where city information is needed.

[tarleb]

The suggestion is fueled partly by the hope that this would make it easier to find an institutions ROR in case that's desired in the future. It's also a part of institution metadata in JATS and supported by the lens reader (but is required in neither). Now that I think about it, I agree that these are rather weak arguments.

[danielskatz]

Is there a way we could support ROR And use that as part of the generate pdf process, I wonder?

[arfon]

Slightly tangential, but how would people feel about using scholarly-metadata.lua to give authors more flexibility when entering metadata?

I'm definitely open to something like this. Perhaps we should open a separate thread for this discussion as part of our broader efforts to enhance the article metadata for JOSS articles?