Intended for healthcare professionals

Editorials

Authorship is dying: long live contributorship

BMJ 1997; 315 doi: https://doi.org/10.1136/bmj.315.7110.696 (Published 20 September 1997) Cite this as: BMJ 1997;315:696

The BMJ will publish lists of contributors and guarantors to original articles

  1. Richard Smith, Editor
  1. BMJ

In April this year we suggested that the concept of authorship in science was so broken that it should be scrapped and replaced by something different.1 Instead of authors there should be contributors and guarantors. Since then the debate has begun to motor. The originator of the idea of contributors—Drummond Rennie, deputy editor (west) of JAMA—has spelt out the case for change in detail.2 The Lancet has adopted the system.3 Many people have written to us about the subject, and most favour change: today we publish 16 of their letters (p 744).4 We propose that, where the authors want it, we will publish lists of contributors and guarantors with papers describing original research. When authors choose not to do this, however, we will continue to publish lists of authors in the traditional way.

We will need to experiment with exactly how the new system will work. Rennie et al suggest that all contributors should meet and agree a description of who did what.2 This might be detailed—as in the paper by Rennie and others—or brief, as has been the case with most Lancet papers. The optimum probably lies between the two. One principle is that somebody should accept credit and accountability for every part of the process, including having the idea, undertaking a literature search, design, collecting and analysing the data, interpreting the results, and writing the paper.

Contributors will also have to decide where to draw the line between contributors and those who will be acknowledged, but we join Rennie et al in suggesting that contributors should include all those who “have added usefully to the work.”2 They might include somebody who suggested the idea and design for the study but did nothing further, or somebody who collected many of the data but was not concerned with design or analysis. Rennie et al suggest that contributors should agree on the relative size of their contributions to decide on the order in which they will be listed. Those who have contributed most will come first. We are not convinced that this is necessary but will be happy to be guided by the experiences of contributors. If researchers want to be listed in order of size of contribution we will do so, but we won't insist on it. The current Vancouver guidelines point out that readers should infer nothing from the order of authors since conventions differ.5 This warning should probably still apply to lists of contributors, at least for now.

Contributors will be fully responsible for their contribution, but at least one person—the guarantor—needs to accept accountability for the whole work. “Guarantors,” say Rennie et al, “are those people who have contributed substantially, but who have also made added efforts to ensure the integrity of the entire project. They organise, oversee, and double check and must be prepared to be accountable for all parts of the completed manuscript, before and after publication.”2 The Lancet has not adopted a system of guarantors, but we agree with Rennie et al that at least one person must take overall responsibility. There have been too many cases of fraudulent research where nobody accepts responsibility. But we are not convinced that guarantors have to have double checked every aspect of the research. Will this be possible if the study includes molecular biology, statistics, and economics? More controversially, does the guarantor necessarily have to have contributed substantially? Mightn't the guarantor sometimes be the person who hired all the researchers even if he or she hasn't contributed much to the particular study? In other words, guarantorship might be akin to ministerial responsibility. The role of the guarantor will become most important when there are serious questions about the integrity of the research. One thing a guarantor might therefore want to do, and which in turn will lend strength to the role, is to satisfy himself or herself that standards of research ethics and supervision are high in the departments where the contributors come from.

In moving from authors to contributors and guarantors we are entering a new era, and it seems wise not to be too prescriptive. We need to learn from experience and adapt the new system. We look forward to working with contributors to improve both the credit and the accountability of published research.

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
View Abstract