Saturday, January 8, 2011

Scientific scandal

A colleague of mine has recently posted an excellent blog regarding the recent scandal arising out of the putative relationship between vaccinations and autism. Public views about science vary a lot but there are those who would point to this scandal (and others) to support an anti-science viewpoint. They might conclude that by being less than 100% perfect, science loses all credibility. Those holding such a 'holier than thou' position should perhaps examine their own lives and consider how perfect their behavior has been. Anyway, I have some comments regarding my colleague's blog:

Lest a gratuitous co-author (i.e., someone insisting on being added to the author list, or accepting an invitation to be a co-author on a paper about which they know little or nothing) choose to deny any responsibility for the content of any paper on which they are a co-author, they should realize that if they have any positive scientific reputation, they're lending their credibility to that paper. If they're co-authors on a fraudulent paper that gets published, they'll inevitably suffer a significant loss of credibility among their peers. No amount of rationalizing can wash their reputation fully clean after that.

That said, however, there has to be an element of trust in a scientific partnership. It's an ideal, but for the most part, individual scientists on a paper with multiple authors don't actually review and replicate everything in every paper. That's simply not a realistic obligation. Nor is it plausible to add a "statement of work" for each author to the paper. In the real world of scientific collaboration, individual authors bring different skills and efforts to the whole, and individual scientists legitimately may have only a superficial understanding of the specialty knowledge brought to the project by their collaborators. If don't trust your colleagues, you'd have to know everything about everything and do everything yourself. Why bother with collaborators in that case?

Funny thing about trust: it can be hard to gain and easy to lose, but as I see it, the default position should be to trust until something happens to violate that. If you trust no one until they've proven their trustworthiness in some sort of a crucible, you'll have a very limited set of colleagues. Anyone can be victimized by a collaborator who turns out to be untrustworthy. It's not fair that some scientist would take a serious hit to their credibility if they were so victimized, but the world isn't necessarily fair! You only have control over things that you do, so it behooves any scientist to behave with integrity, irrespective of what anyone else does.

It's unfortunate that science has such scandals, but it simply reflects the undeniable fact that science is a human endeavor, with all that that implies. Not everyone will always choose to abide by its standards. Science as a whole isn't tarnished by the scandals perpetrated by the deeds of an unethical few. But as my colleague has said, we all feel a collective revulsion and sadness when such things come to light. And there are those who will seize on such events to attempt to discredit science as a whole.

The scandal does underscore the risk you take as a "gratuitous" co-author, and I believe, like my colleague, that the benefit of adding another paper to your CV doesn't outweigh that risk. I always trust my collaborators to do their work with high integrity and I'd be shocked and ashamed to find that a co-author had done something so contemptible. But there's always a finite possibility it could happen to me. If something like this happened to some scientist I know, then my default assumption would be they weren't truly responsible for the ethical lapse of another. This doesn't contradict with my assertion that they're formally responsible for the content of a paper on which they're a co-author, however. Real life can be complicated ...

3 comments:

===== Roger ===== said...

Chuck, thanks for the compliments and insightful additional discussion.

I agree fully that there must be some modicum of trust in a scientific relationship. That said, I'll rehash an old Russian proverb used (ironically) by Ronald Reagan in Cold War negotiations with the Soviets: trust but verify. This applies not only to the improbable but nonzero potential for deliberate misconduct (which we dearly hope never will occur among our own colleagues), but the far more common problem of simple human errors in calculation, interpretation, arrangement or logic. Diligent, pre-submission proofreading and (to the extent reasonably possible) verification of the collaborated work isn't a matter of mistrust, but instead, a part of the entire process (along with later formal review/editing) of making the paper the best it can be. It's a noble and needed duty of every named author!

If nothing else, it lessens the risk of corrigenda or corrective third-party commentary, post-publication. I've been involved on the delivering end of such Comments, and the entire time, I had the nagging notion that more thorough internal review by a particular co-author would have prevented the entire ordeal. That co-author later confirmed it privately.

You're right, as co-authors we can't be intricately familiar with *everything* done by every other author on most papers; but we certainly can proof their words and work to the best of our abilities. I know first-hand that you do. Instead of resenting that, I appreciate it! The Lord and Doswell each know that I'm prone to occasional dumb mistakes or faulty extensions of logic; so the more closely a fellow author can pick nits from a paper with which I am affiliated, the better!

Familiarity breeds either trust or contempt, in life as in science; and we hope that it's the former when we're writing papers with colleagues we've either known for some time, or who have reputations for integrity.

As you and the BMJ editorial each allude, the element of trust can't be viable when authors are signed onto a paper for frivolous, gratuitous reasons. That's why I hammer that co-authorship angle so hard. If a gratuitous co-author gets in deep trouble for the crap to which he/she is associated, that's well-deserved trouble IMHO.

El Gran Rogelio said...

Chuck, thanks for the compliments and insightful additional discussion.

I agree fully that there must be some modicum of trust in a scientific relationship. That said, I'll rehash an old Russian proverb used (ironically) by Ronald Reagan in Cold War negotiations with the Soviets: trust but verify. This applies not only to the improbable but nonzero potential for deliberate misconduct (which we dearly hope never will occur among our own colleagues), but the far more common problem of simple human errors in calculation, interpretation, arrangement or logic. Diligent, pre-submission proofreading and (to the extent reasonably possible) verification of the collaborated work isn't a matter of mistrust, but instead, a part of the entire process (along with later formal review/editing) of making the paper the best it can be. It's a noble and needed duty of every named author!

If nothing else, it lessens the risk of corrigenda or corrective third-party commentary, post-publication. I've been involved on the delivering end of such Comments, and the entire time, I had the nagging notion that more thorough internal review by a particular co-author would have prevented the entire ordeal. That co-author later confirmed it privately.

You're right, as co-authors we can't be intricately familiar with *everything* done by every other author on most papers; but we certainly can proof their words and work to the best of our abilities. I know first-hand that you do. Instead of resenting that, I appreciate it! The Lord and Doswell each know that I'm prone to occasional dumb mistakes or faulty extensions of logic; so the more closely a fellow author can pick nits from a paper with which I am affiliated, the better!

Familiarity breeds either trust or contempt, in life as in science; and we hope that it's the former when we're writing papers with colleagues we've either known for some time, or who have reputations for integrity.

As you and the BMJ editorial each allude, the element of trust can't be viable when authors are signed onto a paper for frivolous, gratuitous reasons. That's why I hammer that co-authorship angle so hard. If a gratuitous co-author gets in deep trouble for the crap to which he/she is associated, that's well-deserved trouble IMHO.

Chuck Doswell said...

Rogelio,

"Doswell" has no Lord-like qualities. I also need someone to catch my errors, which are far from rare. As you know so well.