We would wager handsomely that all of you reading this can think of someone who has been named as an author, but did not strictly meet all of the above criteria. We would also wager, although perhaps not so handsomely, that you can think of someone who could have been listed as an author, but was not. Does any of this matter?
   Of course it does. Publishing articles is like getting a certificate from a school. It’s proof that you have done something. The better the school (i.e. journal), the more kudos you acquire. Such kudos (and visibility) is likely to lead to further research funding, a better job, or invitations to participate in conferences or commercially-funded research and product advocacy. Looking beyond personal gain, being named as an author means that you stake your claim to be part of a scientific community; your colleagues will want to compare notes at meetings, over the phone or in the corridor. The stakes are high, so anything that interferes with the process of attributing authorship must be examined.
   It is often politic to name one’s senior as an author. The senior might not have made ‘substantial contributions’, ‘drafted the article’ or given ‘final approval’, but if you don’t scratch his or her back, he or she won’t scratch yours. Looking beyond personal politics, adding a senior’s name, especially if he or she is well-known in the field, may give your article an easier ride through peer review—because reviewers can be less harsh on respected names—or make your article more noticeable once published.
   This practice is prevalent: one study found that 60% of senior researchers at an institution had accepted authorship in this manner.2 This practice is known as ‘gifting authorship’ and we think it is wrong.
   If you were not involved in the research, why should you have your name on the article? If you cannot truly say ‘I did this’, then why lie? What if the methods were flawed? What if a colleague asks detailed questions about the study? Worse still, what if the data are found to be fabricated? Claiming to have done something that you did not is lying. Saying it is part of scientific culture—or, more accurately, the politics of scientific culture—does not excuse it. Do we want the scientific record to be filled with lies? And if we can’t trust the authorship, can we trust the findings in the article?
   There is also a problem with authors not being listed. All those that meet the above criteria should be listed but sometimes they are not—these authors are called ‘ghosts’. Clearly it is unfair for an individual not to receive credit for his or her work. On that premise alone, ghosting is unforgivable. However, there are also times when authors—or the funder of the research—would prefer to keep someone hidden. Any respectable journal should ask each author for his or her conflicts of interest. Knowing such information helps reviewers, editors and readers to understand the authors’ motivations. Much attention is paid to the idea that people being paid by pharmaceutical companies may submit biased work, but our experience suggests that personal and scientific beliefs are at times equally troublesome. However, if a bona fide author has not been listed, how can the reviewer, editor or reader take their motivations into account?
   By not being listed as an author, the individual is saying ‘I do not meet the criteria for authorship’. If that is not true, then the individual is lying. By its very nature, it is hard to quantify ghost authorship, but there are estimates that 10% of manuscripts are written by ghosts.3,4So we ask again: Do we want the scientific record to be filled with lies? If we cannot trust the authorship, can we trust the findings of the article?
   There are times when someone has made a contribution to an article, but has not fulfilled all of the above criteria. For instance, a technician may have helped with the design and acquisition of the data, or a statistician may have helped with the conception, design, analysis and drafting. Such people should be acknowledged and, as part of their acknowledgement statement, their conflicts of interest should also be mentioned.
   The above criteria for authorship are often questioned, and perhaps rightly so. Who said that those editors can tell the scientific world when they can and cannot be named on their own work? Whether someone should be an author is often highly emotional—people feel they deserve some public recognition for their contribution. This disquiet—and the general difficulties of applying the above criteria—has given rise to the idea of contributorship.5,6Each contributor to an article has to explicitly declare what he or she did. How this is achieved varies: some journals ask authors to explain their contribution in their own words,7 while others have a checklist.8
   There are clear advantages to this approach. Now someone can be recognized as an ‘author’ by explicitly saying what he or she did. Their work can be recognized even if they are not in a position to give ‘final approval’ to the article. It means that authors need only be held accountable for what they did, which has been openly declared in the contribution statement. If a senior colleague wishes to be listed as an author simply because he or she is the boss, then their contribution statement should make the limited nature of their involvement transparent and their egotism plain. One caveat, however, is that one author must take full responsibility for the article; such authors are called the ‘guarantors’. Despite the flexibility of the contributorship approach, some journals still vet whether contributions justify authorship, as they believe that they are avoiding the dilution of the meaning of authorship (the contributors of this article are divided as to the value of this approach).
   Research is changing; hence authorship also needs to change. Increasing division of labour and multi-institutional projects in recent years challenge how to list authors and who takes overall credit. Given the virtually unlimited availability of space on the internet, and the increasingly acceptable idea of transparency in public life, we do not see why an article cannot have hundreds of authors, each with their contribution statements and conflicts of interest published as part of an article. The aim should be to accurately and fully convey who deserves credit and why, and who has responsibility for the integrity of the work.
   We would like to end with some advice for authors: