At last week’s Charleston Conference, I was asked to participate in a panel discussing the topic of trust — how it’s created, maintained, and potentially weakened. The panel was constructed so that there was a publisher, a researcher, a librarian, and a vendor, with each providing his or her perspective on the matter. I learned a lot from the other speakers, but was asked after the conference by a few people to write out the gist of my presentation as a post here.

Here is the written version of the talk.

I think there is a range to trust — trust isn’t binary, a simple comparison between trust and distrust. At one extreme, there are people and information sources I would trust with my life. At the other extreme, there are people I wouldn’t trust as far as I could throw them. And in between, there are sources and people who I trust to a point, but they’ll require backup before I really trust what they say. This is the famous, “Trust, but verify,” stance President Reagan took in the 1980s.

You end up with a spectrum of Throw – Verify – Life.

Along this spectrum, where do the participants in bringing research to light fall?

• Authors — While it’s politically correct to say that “we trust our authors,” the fact is that we hold this statement over their heads like the proverbial Sword of Damocles. Meanwhile, we behave as if we distrust them, a state created by being burned so many times before by author misadventures. So, we require disclosure of commercial interests and attestation to actual authorship. These are not things you require of people you trust. Consequently, authors are generally somewhere between the “Throw” and “Verify” points of the spectrum I’m using.
• Sponsors — Commercial sponsors are clearly the most problematic and more toward the “Throw” end of the spectrum, but even apparently philanthropic sponsors deserve some scrutiny. They can have political or social agendas driving their funding with indirect conflicts or shadow funding driving their choices. Therefore, I put sponsors just to the left of authors, a little more toward the “Throw” side of the spectrum.
• Editors — Editors are an interesting bunch for publishers. We trust them, but also know that their tenures are unpredictable and finite, while the identity of their role is more permanent. Therefore, we watch them a bit. Most are fantastic partners in brand management, editorial advancement, and business propriety, but there have been cases of editors going off the reservation, so to speak, and abusing their positions for the sake of cronyism or personal egotism. I put editors closer to the “Verify” line, but not across it entirely.
• Peer Reviewers — Authors, especially spurned authors, often show distrust of peer reviewers and even the process they are part of. Publishers and editors also trust peer reviewers only so far — if two disagree, we seek a third or probe further or both. We ask them to disclose conflicts, to recuse themselves when appropriate, and so forth. The trust in peer reviewers is about the same as it is for authors.

Aside from these four main players, there are journalists and policy makers, both of which lie again between the “Throw” and “Verify” points on the spectrum.

So, if there is so much distrust in generating any particular piece of scholarly content, how is it that we trust so much of it? Because distrust is part of testing materials for trustworthiness — skepticism is inherent to science. Taking that further, we trust the process writ large. That is, we trust our distrust to a high degree. And we trust the brands that are emblematic of this process, especially those that do it particularly well and reliably.

Therefore, the risk to trust lies at two points — brand and process. And this is where my concern about how we’re conducting ourselves comes in. Are we careful stewards of our brands? Are we managing the underlying process while remaining conscious of the implicit promise we have with our audiences and ultimately with the public?

Can one shift too many lead to a precipitous loss of trust?

This isn’t fanciful thinking. Go out into the “real world,” and listen to the fatigue and cynicism around new scientific findings among the general public. Another study. Another finding. Fish is heart-healthy. Mercury in fish will kill you. Texting while driving is dangerous. States that ban texting see no decrease in traffic fatalities. What’s to be believed? Combine this much more open battle of scientific findings with the pseudoscience being offered by financially or politically motivated entities, and the realm of trust around science is already fairly polluted.

Yet, we continue to put strong brands representing an understood process on top of new ventures with variant processes. And this strikes me as risky in a grand sense.

PLoS has done it, placing the PLoS brand atop two traditional journals along with another non-specific journal with different acceptance criteria and publishing practices.

BMJ is doing it, stretching the BMJ brand to encompass not only BMJ but also books, a “cases” journal, and an OA journal, each with a different process for publication.

Nature teeters on the edge, publishing specialty journals along with Nature Communications, which uses traditional peer-review but has different acceptance criteria.

This brings us to the vaunted “Spandex Principle,” which states simply, “Just because you can, doesn’t mean you should.” (Props to MB for reminding me of this important principle.)

At the same time that brands are stretching to encompass more ways to publish, thereby stretching their promises and credibility, the literature itself is expanding. More papers are being generated, more of those are being accepted — yet, a lower percentage are being cited. So, the trend itself is for researchers to ignore more of the literature. Yet we continue to push more into the system.

One of the newer criteria being accepted as valid is the “methodologically sound” criteria. Phil Davis recently posted about this, while I, in a discussion at another meeting, heard a prominent scientist admit that even this seemingly objective measure is full of subjective criteria — interest, novelty, appropriateness, and so forth. Yet, to rationalize publishing more papers, usually in an author-pays model, some publishers trot this out as if it’s Kevlar when it’s really just more Spandex.

We are stretching the genre of scientific communication to the point of absurdity. On local levels, brands are being used to dress this expansion in what seems a comfortable and compliant fabric, but from the outside reveals unpleasant bulges and irregularities. Overall, the effect is one of rapacious consumption and boundless excess.

We can’t be self-congratulatory about where we’re taking publishing right now. We’re making money feeding the “publish or perish” machines at universities, which are being fed by friendly funding entities, which are subject to political whims. Those whims are about to change, and after the proper period of latency, the level of output (from the UK, the US, and other economies) will likely fall. Is this the only remaining filter on scientific communication — a gross funding filter? Or are we playing our part in filtering information, cementing trust with the public, ensuring a correct process, and building brands that have fidelity to their promises?

If we continue to let the process drift and our brands stretch, we might realize that even Spandex has load-bearing tolerances. When that seam splits, I don’t want to be anywhere in the vicinity.

And if we drift from earned trust — trust generated by skepticism overcome — to assumed trust, then we are asserting something different. We are asserting faith.