A few weeks back I saw an OA paper published in PNAS that has over 37,000 authors. (Well, that’s one way to defray the costs of OA charges!). There are ten regular authors, plus “EteRNA Participants” and a link in the footnotes to the EteRNA author list delivered as a 2Mb supplementary CSV file (ht Greg Jordan).
Science magazine commented:
researchers have now crowdsourced their experiments by connecting players of a video game to an actual biochemistry lab.
This got me to pondering about authors and OA. I imagine that no-one else in the world is interested in this, but here goes anyway.
Deciding who should and who should not qualify as an author has received a good deal of attention over the years, including a post by me. The question of what address each author should use is much less scrutinised. In particular, I have not seen generalised guidance on whether the address given should reflect the address at which the research was carried out or the current address of the researcher.
It seems fairly obvious to me that the ideal is to always list the address at which the research was carried out, with a footnote listing the current address if that is different. But perhaps that reflects my primary interest in things like attribution of credit. For me the address (or affiliation) is there partly to show which institution “owns” the credit for the research, and partly in order to make it possible to communicate with the author. But I have come to realise that many authors are not concerned with that first function. One said to me recently “I put on my new address as my old email doesn’t work any longer and I wanted people to be able to contact me here”. The idea of putting a research address AND a correspondence address had not occurred to him.
I am forever looking at lists of publications and trying to decide whether they should or should not be included in the list of outputs from this Institute. We search for anything that mentions the Institute in the address field, and add those to our outputs database. This requires some vigilance though as the search throws up many papers that have our address in but turn out to be authored by current staff before they came here. Only rarely is this clear from the addresses given on the paper.
Does this matter? I think it does for two reasons. One is that mis-attribution to Institution X rather than Institution Y can potentially affect bibliometric analyses. OK, maybe that is a marginal effect (a guesstimate would say 10% of papers include a wrong address like this) but who knows? The other reason is that it can confound the picture around Open Access compliance. We have been set a target for compliance as an Institute and there may be a financial penalty if we do not make it. I am not sure how the compliance calculation will be done, but if it involves a simple search for our address then the denominator in the calculation will be higher than it should be, probably making our compliance appear lower than it really is (depending on the open access status of the extra papers).
Guidance to authors is patchy. I have not made a detailed study of journal policies, but I found a few encouraging signs. One Elsevier journal stipulates:
If an author has moved since the work described in the article was done, or was visiting at the time, a ‘Present address’ (or ‘Permanent address’) may be indicated as a footnote to that author’s name. The address at which the author actually did the work must be retained as the main, affiliation address. Superscript Arabic numerals are used for such footnotes.
- the names of all authors (first name, middle initial, last name) and their departmental and institutional affiliations at the time the research was done. Indicate which authors are associated with which institutions by listing the appropriate author initials in parentheses after each affiliation listed.
- If an author has changed affiliations and wants this information in the article, then this information should be included in a separate line on the title page.
But Nature says only:
ensure addresses and affiliations are current
I was pleased to see that the issue had been discussed briefly in a thread at ResearchGate, though opinions varied. The actual question posed was what affiliation should be reported if the experimental work was carried out at Institution X but the data analysis and writing up was done at Institution Y, which is more tricky. I think it is justified to use both addresses in that case, provided the work at Institution Y was ‘substantial’. Another tricky example is where a review article was started in one place but finished in another. Again, probably both addresses are justified.
I am not about to start a big campaign about this, but maybe someone will notice and slowly more people will adopt the idea of using a separate Research address and Correspondence address. I did see that ORCID are going to launch an affiliation module, so perhaps they might help to spread the word.
Thanks for reading this far. I feel better now I have got that off my chest!
I have mixed feelings about this. Sometimes an institute tries to “own” something they had very little to do with. If a first author is a recently-made-redundant postdoc who has moved on to a place that actually supports them, I can see why they feel tempted to bury the past in their affiliations.
Jenny – I agree that there may be a host of different reasons in individual cases. The ones I see are mostly papers with a few authors from a particular institution, and one from ours. When I check I see that our author previously worked at that other institution, and in most cases I am told the work was not done at our place. In those circumstances, I think our claiming it would not be correct.
There’s another issue here to do with email addresses. I can’t comment on PubMed commons because the email addresses I had were all institutional ones to which I no longer have access!
Hmm. Seems like a flaw. Is it possible to use ORCID instead?
Might be easier if institutions weren’t so quick to shut down emails of postdocs and PhDs – I’ve been lucky to get a couple of extensions but seriously what would it cost.
This really has more to do with your earlier post
http://occamstypewriter.org/trading-knowledge/2012/10/05/authorship/
but I humbly suggest that most of the authors on those multi-thousand-page papers shouldn’t really count as authors – acknowledged, certainly, but not authors. The question is how many of those authors would feel able to stand up and answer, hand on heart, that they grasped the entirety of the study, or had any major part in its design or execution?
Henry – I think I rather agree. Just imagine circulating the early draft to 2000 authors and getting all their responses back. Then circulating another draft, revised in light of those responses, to all 2000 again for more responses. I think it would be an impossible task to ensure that all authors were happy with and could take responsibility for the whole of the text.