Should we mind if authorship is falsified?

In a typically thought-provoking piece, Louis Menand asks, “Should we mind if a book is a hoax?” In his article, Menand (whose father taught the best course I ever took at MIT, in which we learned that eternal vigilance is the price of liberty) focuses on imaginative literature written by white people but attributed to ethnic minorities. Or, more generally, comfortable people writing in the voices of the less comfortable (thus including, for example, fake Holocaust memoirists who make up their life stories but not their ancestries). His take on it is that there’s an oversupply of well-connected white folks who can pull off the conventions of literary writing, along with an unsatisfied demand for literature sharing the experience of people who’ve suffered. Put together the supply and demand and you get a black market.

Reading Menand’s article made me wonder if there’s anything similarly going on with scientific or scholarly writing. We do sometimes see plagiarism, but that’s more about taking credit for someone else’s work—plagiarism is what lazy and greedy people do—whereas Menand is talking about the opposite, people who do the work but don’t take the credit.

Does authorship matter at all?

For a scientific or scholarly article, what does verifiable authorship get you, the reader or consumer of research? A few things. In no particular order:

– The author is a real person who stands by the work and is thus using his or her reputation as a sort of collateral. In some sense, this works even when it doesn’t work: consider names such as Hauser, Bem, or Wansink where, at first the reputation bolstered the work’s believability, but then the weakness of the published work dragged down the reputation. Reputational inference goes both ways; recall the Lancet, the medical journal that’s published so many problematic papers that publication there can be a bit of a warning sign—maybe not so much as with PNAS or Psychological Science, but it’s a factor.

– Data and meta-data, description of experimental protocols, etc. There’s a real-life person and you can go to the real-life lab.

– Information about the authors can give a paper some street-cred. For example, remember that paper claiming that single women were 20 percentage points more likely to support Barack Obama during certain times of the month? That paper had both male and female authors. If all the authors were male, I wonder if it would’ve been considered too silly or too offensive to publish or to promote.

– Responsibility for errors. Sometimes a paper is presented as single-authored even though it is clearly the work of many people. When there’s an error, who’s to blame. It should be the author’s responsibility, but perhaps the error occurred in a part of the paper that the author did not actually write? It’s hard to know.

Summary

In the above discussion I’m purposely not considering issues of fairness, scholarly due process, etc. Setting all that aside, my focus here is on the way that falsification of authorship can directly reduce the usefulness of a published work of scholarship.

Remember how Basbøll and I discussed plagiarism as a statistical crime, based on the idea that plagiarism hides important information regarding the source and context of the copied work in its original form, information which can dramatically alter the statistical inferences made about the work?

Here I’m saying that this concern is more general, not just with plagiarism but with any misrepresentation of data and metadata, which includes authorship as well as details of how an experiment was carried out, what steps were done in data processing and analysis, and so on.