When Pigs Fly — Changing author lists on scientific papers

A sizeable minority of researchers favour changing the way which author contributions are presented on scientific publications. A twitter poll receiving over 1000 responses found that although 62% of participants back the current system, where names of lead authors appear first and corresponding authors last, there is also significant appetite for change.

So is it conceivable that we may move to a better system?

Ah…

CRediT where credit is due

Colleagues of mine who come from press or tech backgrounds are often bemused by scientists’ obsession with ensuring all contributions to a project are properly attributed. But the reason is simple, receiving recognition for work done is vital for a scientist’s career. For junior researchers’ it is necessary to obtain the next job and for established academics it can help land the next grant. The current system is simple and easy to understand, but it is also limited by the fact that it is simple and easy to misinterpret.

To give an example the following screenshot shows how a paper is displayed at eLife, but the principles are the same for most journals.

We can see the title and below it authors names are listed in order of importance, the first author is deemed to have done most work, and the last (corresponding) author oversaw the research and obtained funding.

However, little attention is given to the names in between.

In this instance, the second author has done all the bioinformatic analysis on the paper, but in a standard layout, you have no way of realising (or proving) this. Author order is too simple to reliably determine an individual’s contribution and often undervalues the work of co-authors.

Susana Sauret-Gueto — Current system is fair for a) one main author papers, but not for b) team papers. Thus, it does not help b)

In a small step to address this oversight, many journals have recently adopted a system for standardisation of author acknowledgements known as CRediT. This is a means of clearly stating what each author contributed to a manuscript.

These contributions are displayed in different ways at different journals, but following on from the previous example, they can be viewed at the click of a button (see below)

Contribution statements can be viewed as a drop down list

In this drop down menu you can see that the second author Sara, contributed to the bioinformatics.

A politician’s fallacy?

So far so good. CRediT does in part help to show what each author has done for a manuscript, and it would appear from twitter that researchers are overwhelmingly in favour of using the system. So why even consider doing more?

Well, after many discussions with early-career researchers there is a sense that even with CRediT the current system does not do enough to recognise their contributions to papers.

Paul Agapow — Is there a need for change to way author contributions are recognised? — Hell yes. Review panels treat author lists like reading tea leaves

There is still an over reliance on first author papers in job applications and arguments over author placing remain common. Further as Pietro Gatti found previously many appear to be indifferent or disillusioned with the current format of contribution statements.

Has the issue been addressed at all?

Going one step further

A more radical solution would be to try and increase the importance of the contribution statement by replacing the conventional hierarchical arrangement of author lists with an alternative. I therefore asked the twittersphere if they would back changing the system to display contributions alongside an alphabetical arrangement

Would you support replacing author order on papers with alphabetical listing and displaying contributions? 1084 votes, 62% — No, 31% — Yes, 7% — Other

The response was passionate and mixed. Unfortunately much of this focused on my unwise suggestion of alphabetical ordering as a solution (which was quickly debunked by many people pointing to the fact there has been a lot of research pointing to an unfair advantage for those at the top of the list). Randomisation was suggested as an alternative

D. Weigel — Alphabetical has it’s own issues. Random order preferable over alphabetical. Arranged in a circular, not linear manner.

but perhaps most interestingly a number of key concerns were raised that highlight short comings with current acknowledgement statements and pin pointed key issues any replacement would need to address —

1. Ensure the relative contributions of authors are properly acknowledged

The most common concern was that any alternative system must not go from overestimating to underestimating the amount of work key individuals have made to a paper. There were also worries that only the most recognisable names would be credited with a particular article, and it would disadvantage minorities.

Elisabeth Bik — In biology, often 1 or 2 persons did most of the work. Would be sad if those would end up in the middle.

Khalid K. Alam — As someone with upcoming papers & a career at stake, I say no. Listing contributions is important but order says alot.

Current author contribution statements give no idea of how much importance should be placed on them and it is also difficult to get an immediate overview of who contributed what. A simple solution which is worth further consideration came from Richard Edwards:

Richard Edwards — I think a contribution grading system, (“primary”, “major”, “minor”) would be better with classifications in the contribution statement.

An excellent point was raised by Kostas Vavitsas which actually highlights a potential positive effect of implementing change

Kostas Vavitsas — Then we’d need to redefine productivity for #ECRs eg overall quality of all work instead of one or two big papers that are one’s babies

would redefining productivity help with a culture change to lessen some of the negative effects of a hyper-competitive research environment.

2. Dealing with large collaborative projects

There could be an issue with recognising author contributions on massive collaborative projects. Robert Garisto pointed to the solution currently employed in high-energy particle physics —

Robert Garisto (HE Physics) — we allow any order authors want & footnotes to express contributions.

3. Referencing of papers would need to be changed

Moving away from author order would require a radical change to the way papers are cited.

Christian Hardtke — A radical redesign of citations would be needed. e.g. emphasis on paper title instead if 1st author

one suggestion was to replace names in references with the DOI:

Sequential neuromodulation of Hebbian plasticity offers mechanism for effective reward-based navigation. Z Brzosko, S Zannone, W Schultz, C Clopath, O Paulsen (2017) eLife 6:e27756

which could become -

Sequential neuromodulation of Hebbian plasticity offers mechanism for effective reward-based navigation. DOI:10.7554/eLife.27756

4. Change cannot be done alone

A potential killer of any initiative was the observation was that any change in referencing would likely need to be universally adopted to ensure consistency:

Magnus Nordborg — the problem is the usual one: one journal cannot do it alone

Mick Watson — Changes like this cool but need to be rolled out across all journals to work

For a full implementation this would definitely be true. It would be impossible to have a system where some journals have references based on author order an others do not. However, it might be conceivable to maintain two systems for an interim period, whereby traditional author lists are also provided by a journal for external use, and the option to employ a non-hierarchical system could be offered internally.

A potential solution

The essence of what I wanted to suggest was more clearly stated by others

Swapnil Bhatia — the proposal is to REPLACE reliance on author order with a DIFFERENT contributions “data structure”.

Laurent Gatto — There’s remembering and assuming 1st author was major contributor. Goal is to dissociate 1-dimensional order & multivariable contributions

and Swapnil Bhatia suggested an alternative to alphabetical ordering which could help address many of the concerns raised:

Swapnil Bhatia — This is a good idea. I’ve advocated this myself before, inspired by movie credits.

what might this actually look like? I have done a mock up below with one possible option.

There are several key aspects to note: (1) It is easy to get a quick overview of contributions to each area (2) information about the relative contribution could be captured by ordering names under each heading, with some authors contributing more to certain areas than others (e.g. Johan sanz Terre — Software).

Moving forward

I really liked this idea of moving to a ‘film credit’ style of displaying author lists and it will be interesting to discuss further.

But what can be done in the meantime?

For all the early-career researchers that supported changing the current system Sophien Kamoun made an excellent suggestion about what to do to highlight your contributions -

Sophien Kamoun — Author contributions should become formal part of CVs. List what you contributed to paper. Lying will have consequences…

Receiving appropriate recognition for the work put into a paper will remain a contentious issue. The move to implement CRediT is a positive step but more work can and needs to be done to help junior researchers. Hopefully by raising awareness about problems and continuing to discuss what might be done to address we will make a better system.