Conflict of Interest Disclosure

What it is

The requirement attaches to the venue rather than to the person or to the claim. So it travels with the least-read version of the work and falls away from the most-read one. A trade book is therefore the venue where a commercial interest is most consequential to the reader and least likely to be declared.

Nothing in this concept is an allegation of misconduct, and it should never be used as one. Having a commercial interest is not a wrong in itself; it is the ordinary outcome of a scientist whose work reaches a clinic. Authors are not obliged to supply disclosures nobody requires. The object of the criticism is a convention's scope.

The rule that follows is practical. Before citing a trade book's recommendation of any intervention, look up whether the author has a relationship to it, and say so in your own copy whether or not the book did. The check takes a minute and the book will not do it for you.

In effect

The first clean case in this library is The Distracted Mind, and it is clean precisely because nothing improper happened. Adam Gazzaley co-founded Akili Interactive Labs in 2011. The company held the exclusive licence to a university patent on the methodology behind the laboratory game that is the centrepiece of the book's tenth chapter, and he was named on a related pending patent. When the trial was published in Nature in 2013, the paper carried a competing-interests statement naming the company, the role and the patent, because the journal required one. The 2016 book recommends cognitive training in that chapter and carries no disclosure anywhere: not on the copyright page, not in either acknowledgements section, not in the prologue, not on any page of the chapter, not in any of its 357 endnotes, and not in the back-cover biography. The interest was never hidden from the world. It was in the Nature paper, in university press releases and on the author's own website. A reader of the book could not find it, because it is not in the book.

The second instance is a different kind, and it concerns intellectual independence rather than commercial gain. In The Elephant in the Brain, nine citations are to Robin Hanson's own work, and the character of them is the finding rather than the number: of five examined closely, only one is conventionally peer-reviewed, the others being a self-hosted working paper, a journal that did not then operate conventional peer review, a post on the author's own blog cited for a substantive claim, and a second self-hosted working paper. In a bibliography about forty-four per cent refereed, the author's own contributions come almost entirely from the part nobody refereed. Separately, a colleague's then-unpublished manuscript is cited five or six times, marked as a preprint each time, and supplies a large share of the education chapter's figures. That colleague read early drafts, supplies a cover endorsement and is a departmental colleague of one author, none of which is stated in the chapter where his manuscript is doing the heaviest lifting. The chapter's most quotable figure, an estimate that signalling accounts for up to eighty per cent of the value of education, carries no note at all, and critics reconstructing the underlying numbers arrive at a signalling share of roughly eight to thirty-five per cent.

The strongest test is Your Brain on Art, and it sharpens the rule. The book carries no disclosure statement, no conflict-of-interest note and no funding statement anywhere, and it reports throughout on institutions, initiatives and products to which its two authors are attached. The credit is owed first: the rear flap states both authors' institutional roles, two parenthetical clauses in the running text disclose further positions, five relationships are disclosed in the text at the point of use, and the copyright page carries an institutional disclaimer. Judged on in-text disclosure alone the book does more than the first instance did. Where the reader is not told, seven passages are involved, including a quoted endorsement from a research director at a lab one author founded, on a page that does not say so, and an approving description of another researcher's company product in a medical claim with no note of his founding interest.

That instance produces the sharpened rule. A disclosure that is not adjacent to the claim is not a disclosure. Front matter, a flap, an acknowledgement or a parenthesis in a different chapter all fail for the same reason: the reader is not holding them at the moment the claim arrives.

What it does not say

It does not rest on research. This is a description of how a publishing convention is scoped and of the instances read end to end here. There is no proposition to replicate.

It does not allege misconduct in any of the cases above. No rule was broken in any of them, every fact recorded is public record, and the criticism is of the convention rather than of any person.

It does not say self-citation or citing colleagues is improper. Academics cite the colleagues they argue with, and a small field is a small field. Independence is what a citation is supposed to supply, and a citation to your own unpublished paper supplies less of it than the typesetting suggests.

It does not only concern commercial interest. The first case asks what an author stands to gain if you believe a claim. The second asks how far a claim travelled before it got back to the author. Both are disclosure gaps and both are properties of the venue.

It does not stop at one interest behind one recommendation. The third case is an authorship arrangement in which the authors' own institutions are among the book's recurring subjects, with the interests declared as biography rather than as interests, which is a difference in kind.

---