Clarifying attribution in a digital world
Understanding the rules
By Mike Taylor Posted on 12 March 2013
The list of authors attached to a paper has provided a comfortable device for apportioning attribution for centuries - not only is it simple and clear, we all have a good idea of what ‘authoring’ implies.
However, with the increasing migration of content to an online environment, the number of ways in which papers can be connected has exploded. We expect references and citations to click through, we want to see what else authors have written, and we are keen to discover who they have collaborated with.
But while this connectivity establishes that there is a link between an author and a paper, it says nothing about the nature of that link. Did the author write the experiment, analyze the data, or were they responsible for running the research program, ie. had very little to do with this particular article?
Bridging the discipline gulf
Depending on your field, you may be thinking about the order in which the authors are assembled, and how you can use this to make sense of their relative status. However, a researcher in another field may interpret that same list quite differently. This is illustrated in the table below in which we take a look at a fictional paper written by Smith, Taylor and Thorisson.
|High Energy Physics||Author list is in alphabetic order, no precedence can be interpreted. Names may include engineers as well as researchers, in this case we could add Bercow, a PhD student who ran the experiment and took care of writing computer algorithms, ensuring the integrity of the data and selecting candidates for trials, etc...|
|Economics, some fieldswithin Social Sciences||Author list is in alphabetic order, no precedence can be interpreted.|
|Life Sciences||Smith the postdoc did most of the experimental work, but Thorisson was the principal investigator who led the scientific direction of the work. The alphabetical order is coincidental.|
|‘Standard’ order||Smith is the senior researcher who did most of the work. Taylor was subordinate to Smith, Thorisson is subordinate to Taylor. The alphabetical order is coincidental.|
Table 1. Varied authorship conventions across disciplines referencing a fictional paper written by Smith, Taylor and Thorisson.
One of the advantages that a reconceptualization of authorship offers us is the proper acknowledgement of work undertaken during the research process without conferring a higher status than is merited. For example, work that is undertaken on computer algorithms would not (at the moment) usually confer authorship by itself – however, a move towards a contributorship model would enable the correct communication of the algorithm creator's contribution.
As publishing opens up, and platforms become more integrated, we are likely to become more exposed to articles that are outside our field, and others that are interdisciplinary. If we can’t rely on the author order to help us make sense of what authorship has meant for a particular paper, can we at least be sure about the class of activity that merits authorship attribution?
With the exception of high energy physics, which will include engineers along with researchers in the authorship list, it’s a surprisingly difficult activity to undertake. There is an extensive and growing literature available covering the ethical dilemmas that arise from attribution.
Understanding the rules
The most widely used authorship rules are known as the ’Vancouver rules’ and were laid down by the International Committee of Medical Journal Editors (ICMJE)(1). The rules, which are followed by several hundred journals – mostly medical - specify that:
“Authorship credit should be based on: 1) substantial contributions to conception and design, acquisition of data, or analysis and interpretation of data; 2) drafting the article or revising it critically for important intellectual content; and 3) final approval of the version to be published. Authors should meet conditions 1, 2, and 3.”
Although these rules are clear and many journals are signed up to follow them, there is evidence that compliance is far from complete, and that the rules are not well understood. In fact, exploring alternative models was the focus of the International Workshop on Contributorship and Scholarly Attribution (2), hosted by IQSS (The Institute for Quantitative Science) at Harvard in May last year. Those of us with experience in publishing would recognize the issues involved in getting hundreds of authors to give final approval to a document: and some fields have published manuscripts where authorship is well into the thousands.
Two phenomena that have become recognized over the last few years are ‘ghost’ authorship and ‘guest’ authorship. Guest authorship occurs when an author’s name is included on the authoring list despite them not having played any part in the research or authoring process. This can come about in a variety of ways: when it is ‘normal lab practice’ to include the head of the lab on all publications; when a researcher has left the organization before the paper is written; and in the hope that a senior and well-respected name will confer additional credibility on a research paper and improve the chances of publication in a high-impact journal. Ghost authorship is a phenomenon that is said to occur when interested parties employ a professional writer, or have staff members on the research team, but don’t include these individuals on the final publications list. The implication is that these people would have a conflict of interest in the outcome of the research (or at least the presentation of the research at publication). By omitting their names, the paper affiliations look more neutral. There is a variety of other literature (3) in this area, alleging political / organizational influence in the creation of authorship lists. There is also some evidence that – when computed – tasks that would have previously conferred authorship no longer have this advantage.
As publishing articles is frequently considered to be the main currency of academic recognition – and is increasingly included in formal rule sets that govern academic status and eligibility for funding – so we can expect an increase in the heat governing this debate. Potentially, if the number of authors included in an article continues to increase, we may also witness a decrease in the value of authorship.
Adopting a contributorship approach
So what about contributorship, and how does it relate to authorship? Authorship is bound to persist into the future, for both ethical and copyright / legal reasons, so contributorship needs to be seen as an extension of existing protocols.
In short, the idea behind contributorship is to disclose what activities the researcher undertook to merit a place on the author list. Research undertaken by the author and colleagues in 2012 indicates that nearly all activities can be classified into a set of between 12 and 15 categories. The first three of these are outlined in Table 2 below. Although the prospect of this additional task as a manual activity would – quite properly – concern all people involved in the authoring and publishing process, this work could be a feature of the various research and collaboration tools (e.g. Mendeley) and services such as ORCID (footnote: ORCID was designed to enable definitions richer than authorship).
|Conceptual and intellectual||Formulation of the research questions; design of the experiments; interpretation of the results; intellectual and moral responsibility for the integrity of the paper, as a whole and for individual contributions.|
|Technical and experimental||Implementation of the investigation and undertaking of experiments; obtaining specimens and subjects; acquisition and processing of data; analysis of results.|
|Organizational and communication||Writing the project proposal and obtaining funding; ensuring compliance, e.g. ethics committees’ approval, informed consent; writing up results into a paper; illustration of the paper – selection and use of data; reporting.|
Table 2. Three of the categories proposed to classify contributorship
Relationships are changing; we are moving towards a richer world with a more detailed and nuanced web of connections. Although the concept of authorship is rooted in our culture and in our minds, contributorship could offer a richer set of definitions, enabling our contributions to human knowledge to be recorded more precisely. The value to us will be in knowing the areas of our expertise and contribution. If the cost is ameliorated through intelligent tools and services, then we can expect to see contributorship becoming one of the hot topics in scholarly communications and publishing.
- Do you agree that contributorship is the way forward? Would it work in your field? Please let us know your thoughts by posting your comment below.
An article on this topic, Fixing authorship – towards a practical model of contributorship, has also appeared in Elsevier’s Research Trends.
TECHNOLOGY RESEARCH SPECIALIST
Mike has worked at Elsevier for 16 years. He has been a research specialist in the Labs group for the last four years, and has been involved with ORCID (and previous projects) throughout that time. Mike's other research interests include altmetrics, contributorship and author networks. Details of his research work can be found on http://labs.elsevier.com.
(1) International Committee of Medical Journal Editors, Uniform Requirements for Manuscripts submitted to Biomedical Journals: Ethical Considerations in the Conduct and Reporting of Research: Authorship and Contributorship.
(2) IWCSA Report (2012). Report on the International Workshop on Contributorship and Scholarly Attribution, May 16, 2012. Harvard University and the Wellcome Trust.
(3) A small selection of the literature available on this topic:
- Honorary and ghost authorship in high impact biomedical journals: a cross sectional survey, BMJ 2011, dx.doi.org/10.1136/bmj.d6128
- Authorship: why not just toss a coin? Am J Physiol Cell Physiol 2008 10.1152/aipcell.00208.2008
- Authorship versus Credit for Participation in Research: A Case Study of Potential Ethical Dilemmas Created by Technical Tools Used by Researchers and Claims for Authorship by their Creators, 10.1197/jamia.M2212
Carl says: March 14, 2013 at 12:54 pm
This is an excellent article, but not objective. Clearly the author is a fan of attribution. Questions not adequately addressed include:
– conflicts in perception and attribution of contribution
– value of ambiguity in complex situations
– value of the exercise vs the cost
Michael Taylor says: March 15, 2013 at 9:06 am
Carl, thank you for your comment. I would say that I am an advocate of contributorship, even if it is to work towards building a better set of commonly understood rules, from which we can derive a definition of authorship.
As someone with a mixed academic background, I quite enjoy notions of ambiguity, and definitely appreciate your first two comments that relate to this. I wrote an article with Dr Thorrison that lightly touched upon this issue: http://www.researchtrends.com/issue-31-november-2012/fixing-authorship-towards-a-practical-model-of-contributorship/ – but it's definitely worth exploring in more detail.
Having done a reasonable amount of work with taxonomies and ontologies, I have often found it very frustrating that the authors of classification systems have built them with a view to guiding the user to the finest possible class, e.g., Sorex araneus, rather than something that all observers could agree on, eg, 'small mousy thing with a pointy nose' – for which there is no formal classification. This is why I would seek to build an evidence-based ontology that can accommodate a degree of ambiguity. Sometimes our best ideas arise in an intellectual space between us, and it's simply not possible to identify a key moment or person who had the idea – hence the need to have flexible definitions such as "had critical intellectual input" without having to dive down deeper.
You raise the question of cost, which is key to this debate. Clearly there are benefits to this approach, clearly people who write papers do not need to spend an extra day or two (or months, in the case of CERN) agreeing definitions. However, we are living in a period where a large number of environments and tools that support the research and authoring process are being developed – and many of these hold relationships between individuals, projects and work. Orcid.org, Mendeley, Sympletic, Figshare, Scopus etc. So my argument would be that the developers of these platforms should consider that we need to be more sophisticated in our representation of these relationships in all parts of our ecosystem, that we can't simply bind together a document and an individual into a bucket labelled authors – since (taken as a whole) we have no working definition of what authorship does. In other areas of our community, the technical environment allows for a more complex and nuanced set of relationships, and I would urge colleagues who work within scholarly tools and platforms to (a) allow for many-to-many individual work relationships, and (b) support the development of a common set of individual work relationships, even if it's at quite a high level.
Michael Orendurff says: March 14, 2013 at 2:25 pm
Another factor that I include is that the author should be able to defend the work in public. If I include the statistician as an author because of a high degree of intellectual content contribution to the paper, I expect her to have enough familiarity with the work in her area of expertise that she would stand up and defend the approach to her peers in a public setting. Guest authors usually cannot perform this key duty.
Michael Taylor says: March 15, 2013 at 8:43 am
Michael, thank you for raising an excellent point. The ability to defend a work is clearly a critical factor in some fields. Equally obviously, there are articles published about some experiments which are too complex to be defended by all the people listed as authors – the papers published from CERN / LHC come to mind, where the number of 'authors' has been known to exceed 3,000. In part, this is because high energy physics will include engineers as well as theoretical researchers as 'authors', but it is also a measure of the complexity of the environment in which they work.
I read some comments on David Calquhoun's blog a while ago (sadly I can't find the link now) where a thread was discussing this – and the reaction of some researchers was to state that these 'authors' aren't what they'd call authors at all. Which is true.
Since we're never going to be able to impose our discipline's rules on other fields – in other words – agree on a common understanding of what authorship is – my hope is that we can, at least, agree on a set of authorship components to which we make attributions – which is what is at the heart of this debate. How we then go about creating a framework that defines principle authors, or legal / copyright authors, is something that can be developed using a common set of definitions.
Ged Ridgway says: March 15, 2013 at 12:59 pm
In this article's definition, "contributorship is to disclose what activities the researcher undertook to merit a place on the author list".
However, some journals use the term rather differently, e.g. "Neurology defines a Contributor as a person who does not meet the criteria for authorship of the study, but who has contributed in other ways, such as collection of data; technical assistance; acquisition of funding; supervision of personnel; contribution of drugs, reagents, equipment, or participants; or editing the manuscript for non-intellectual content."
Personally, I prefer the definition in this article; I think life sciences and medicine underestimate the importance of technical/software contributions, which should merit authorship more often than they seem to, in my opinion.
Jim Goodwin says: March 15, 2013 at 2:51 pm
Attribution is a very complex issue. Author lists and orders are governed by the field, country, and even institution based on tradition, internal politics and reward systems. So, to come up with one size fits all is probably nearly impossible, especially since the different contributions fall under many more than 3 categories (maybe more like tens of categories to fit all cases). What might work better is a paragraph in a section at the end of the paper (perhaps still called Acknowledgements) describing all the authors contributions as well as acknowledgements of the contributions of others more peripherally involved.