This is the second post in an ongoing SciWriteLabs discussion about fact-checking and the appropriate way for writers to interact with sources. Part one, which featured Vincent Racaniello and David Kroll, is here. Today’s participants: Wired senior editor Adam Rogers, who has also worked at Newsweek and was a 2002-2003 Knight Science Journalism Fellow at MIT, and Ivan Oransky, the executive editor of Reuters Health, founder of Embargo Watch, co-founder of Retraction Watch, and an adjunct associate professor at NYU’s Science, Health, and Environmental Reporting Program.
SM: One of the things that struck me from day one of this debate is that people were using phrases that had very different meanings to different people. For instance: I think of “fact-checking” as a somewhat formal process in which someone who isn’t the reporter independently confirms everything that appears in print. This is a fairly traditional glossy magazine conception of the term. Some people seemed to be using it to mean basic editing; in other cases, it seemed to refer to giving sources veto power over any/everything that appears in print.
Adam, in a comment on David Kroll’s initial post, you offered to share Wired‘s fact-check guidelines. Can you give us quick-and-dirty summary of what those are?
And Ivan, when you’re working with students at NYU, do you talk to them at all about what type of pre-pub/pre-post review is appropriate?
IO: I’m all for any practice that helps ensure a story’s accuracy, and I’m firmly against any practice that gives control of copy to a source. The latter is because I think one of journalism’s main contributions to society is keeping people honest — and scientists are among those people.
How does that calculus play out in my approach to fact-checking? What I tell my students at NYU’s SHERP hews very closely to Reuters’ official policy:
Reuters never submits stories, scripts or images to sources to vet before publication. This breaches our independence. We may, of our own volition, check back with a source to verify a quote or to satisfy ourselves about the reliability of factual information but we also need to ensure that in doing so we do not give sources an opportunity to retract or materially alter a quote or information to their advantage.
Interview subjects or their organisations or companies sometimes ask to see the quotes we plan to publish or broadcast before they are issued. We should resist such requests and explain why this is not our policy. If there is no option but to submit quotes for approval, this decision must be taken in consultation with your manager. We should never agree to a quote being materially changed. It is often effective to give the source a tight deadline for approval.
We may agree to “clean up” a quote linguistically, especially when the speaker is not using his or her own language. We do not, however, massage quotes to change meaning. If a source asks you to change substantial information, either drop the quote or do not run the story at all, again in consultation with your manager.
And that policy, in turn, is what I’ve maintained for my staffs at every news organization where I’ve worked.
SM: So I actually find that kind of confusing — and I’m not just talking about the fact that Reuters spells “organizations” with an s instead of a z. What if a writer’s definition of checking with a source “to satisfy ourselves about the reliability of factual information” is to go over the story in its entirety? Does that violate the stricture against “submit[ting]…to vet”?
Here once again I think we get into the whole source/subject distinction. In yesterday’s conversation, Vincent talked about reviewing work for reporters like Rebecca Skloot; what would Reuters say about showing a story to a source who isn’t involved with/quoted in that story?
I’m not trying to be a dick here — this is something that I know comes up. I’ve had instances in my own work where errors would have made it to my copy if I’d limited myself to feeling satisfied with the reliability of factual information — the errors I’m most worried about getting in are the ones that are based on incorrect assumptions.
Adam, I know I’d asked you what Wired‘s fact-checking guidelines are; does it have things it explicitly does not allow writers to do?
AR: First let me say that one of the keys to good fact-checking, and good journalism in general, is transparency and keeping sources of information honest. So the simple existence of Ivan’s Retraction Watch and Embargo Watch makes me really happy.
Seth, you asked for a summary of Wired’s fact-check policies. The basics are: We ask all our writers, for stories of every length, to give us a list of sources with contact information, documentation—copies or URLs—for background material, and transcripts of interviews. That last one is flexible; they have in the past accepted seeing my hand-scrawled notes, for example. They also really like recordings.
The research desk guards the anonymity of sources but still wants to talk to them. Primary sources only—newspapers and magazines don’t count. For legal issues, we need police or court documents, and Conde Nast’s lawyers often get involved. The truth is, we’re not only trying to be accurate but also trying not to be sued.
We don’t show people their quotes, though we do sometimes alter them for clarity—but never changing the meaning or sense.
We do go back to the people being quoted and (without showing them copy) ask them if we’re accurately reflecting their ideas. We also use outside experts to vet what we’re saying.
Editors and writers often talk to fact-checkers before they start calling sources, letting them know about any matters of particular sensitivity. That might mean making sure we’re being sensitive to a source’s emotional well-being, and it might mean making sure that we’re not tipping a source off about the angle of a story.
Things Wired writers aren’t allowed to do? Hmm. They’re not allowed to cite Wikipedia. They’re not allowed to drop a whole book (or 10) on the researcher’s desk and say, “it’s in there somewhere.” But probably they could, if they wanted to, embark on their own fact-check journeys. We tell them that in general it’s a bad idea to show what they’ve written to a source, but also understand that sometimes it might make sense to run particularly technical passages past a source—always with the stipulation that we probably won’t change something for style or sensibility reasons unless it’s demonstrably wrong. It’s also a bad idea to send those parts to sources, I usually add.
IO: Hmm. Seth, I sort of see the confusion, but not really. Maybe this will help: Reading passages over the phone is fine. Showing copy isn’t. If there’s disagreement, and your source isn’t happy with the way things turned out, he or she could post the version you sent, or even use it in a lawsuit. Maybe that’s an unfortunate thing to have to think about, but it’s a real one. I haven’t heard anyone at Reuters say, “Phone OK, email isn’t” explicitly, so I’m interpreting here, but it’s what I tell my students and reporters.
I’m not sure I’m getting the source/subject distinction. If someone is providing information that informs a story, he or she is a source, right?
And Adam: thanks, that’s very kind of you!
SM: Adam, you said, “We also use outside experts to vet what we’re saying.”
Does this mean showing outside experts some stories in their entirety? I suspect that some of the confusion/dissent about this issue stems from just this type of scenario. I can’t imagine anyone looking askance at the fact that Wired goes this extra mile to make sure its stories are accurate. But depending on how that practice is described, couldn’t that be interpreted as vetting stories with sources?
Ivan: you know, I’d actually never thought of it in those terms (the phone/email distinction) but that makes perfect sense. I guess a lot also depends on the relationship with the source.
Finally, the source/subject distinction: To my mind, the hypothetical “outside expert” referenced above would be a source (as would anyone whose knowledge informed anything about the piece); someone who is a focal point of a story is a subject. If I were writing about the recent news about the development of a malaria vaccine, researchers working on that project would be subjects; if I asked a vaccinologist unconnected with that work to tell me whether s/he thought I was correctly interpreting the study, that person would be a source.
I don’t think this is a distinction that is generally agreed on…
AR: Whoo, no. I can’t think of a case where we’d show anyone outside the office an entire story. We very rarely show any piece outside the office.
For what it’s worth, Seth, I draw the same distinction you do. Person the story is about: subject. Person who is consulted for information to be used in the story: source. Subjects can be sources; sources are not subjects.
But now I can’t remember why that’s important. At Wired, at least, we don’t show the story to either of them.
IO: Yes, it definitely depends on your relationship with the source, I’d agree. And I see what you mean about source vs. subject. Your distinction seems perfectly clear and useful; I guess what I would ask is whether that really matters in terms of whether you’d share copy. I’d say no. You have a different relationship with the target of an investigation than with a whistleblower, but I wouldn’t share copy with either.
SM: I get – and share – the knee-jerk reactions both of you seem to have against showing raw or unpublished copy to someone outside of the publishing process. The problem is, when I push myself to articulate why this is, I come up short. There are legal reasons, for sure, but if I’m looking at this from a standpoint of journalistic ethics, I can’t think of an a priori reason why showing your copy to a source sullies the integrity of your story.
There are about a dozen other thoughts I have, but maybe this is a good place to leave things for now…
SciWriteLabs #2: Reuters’ Ivan Oransky and Wired’s Adam Rogers on sources v. subjects & more on the factchecking debate by PLOS Blogs Network, unless otherwise expressly stated, is licensed under a Creative Commons Attribution 4.0 International License.