Meetings of Minds, or, How to Talk to a Scientist


Science writers need scientists. Scientists probably need writers a little bit less, but even so, this relationship is generally productive for everyone involved: Writers get a great story, scientists communicate their work to a broad audience, and readers learn about science. But tension can also creep into the working relationship between writers and scientists. A writer and scientist might have competing views on what aspect of the scientist’s research should be highlighted, for example. Scientists may try to retract or alter quotes after the fact. And sometimes writers get the story wrong, leaving scientists frustrated and distrustful of journalists in general.

I asked three scientists and three science writers to share their thoughts on the scientist-writer relationship and how to avoid these problems. Below is an edited version of the resulting email discussion, which covers how to prevent reporting mishaps (hint: don’t make assumptions), the pros and cons of allowing scientists to review quotes, and the key questions writers should ask.

The scientists participating in the discussion were:

Julio Betancourt, a senior scientist at the United States Geological Survey, an adjunct professor in the University of Arizona’s School of Natural Resources and the Environment, and co-author of two books on environmental history

Karen Lips, an associate professor of biology at the University of Maryland

Daniel Simons, a professor of psychology at the University of Illinois and co-author of The Invisible Gorilla: How Our Intuitions Deceive Us

And the writers were:

Jon Cohen, a staff writer for Science who has written for many other publications and is the author of three books, most recently Almost Chimpanzee: Searching for What Makes Us Human, in Rainforests, Labs, Sanctuaries, and Zoos

Julie Rehmeyer, a freelance science and math writer, contributing editor at Discover, and contributor to The New York Times, The Washington Post, Slate, Science, Wired, and other publications (and a member of TON’s Board of Directors)

Carl Zimmer, a columnist for The New York Times; a contributor to magazines including National Geographic and The Atlantic; and author of 13 books about science, including Parasite Rex and The Tangled Bank: An Introduction to Evolution


Geoff: It’s easy to get scientific details wrong. Writers, have you had any major mishaps because of misunderstandings between yourself and a scientist? And what can writers do when interviewing scientists to avoid inaccuracies?

Julie: The worst cases for me have been for things where it just didn’t occur to me that it might not be true.

I cowrote an obituary for The New York Times. I’ve written about Grothendieck before, and I was somewhat familiar with his work from my time in mathematics—and the latter is exactly what got me in trouble. I was completely certain that his work was (distantly) related to the Poincaré Conjecture (which says that any three-dimensional space without holes is essentially a sphere), and I put that in, because it’s one of the results that regular folks have some chance of having heard of. I was so certain that when I heard some vague complaining on Facebook about that from mathematicians, I thought they were just being overly scrupulous. It took me a while to realize, no, I was just flat-out wrong.

I definitely learned from that one! In that situation again, I would carve out time for explicit fact-checking no matter how tight the deadline. And, of course, it’s a reminder about how it’s possible to be wrong on things that you’re very sure of.

Jon: I made a mistake in February that embarrassed me. It was similar to the problem Julie mentioned: I knew too much for my own good.

The story was an update on the status of Ebola drug trials. I wrote that ZMapp, the famous antibody cocktail, was about to be tested in a randomized, placebo-controlled trial. As I learned, moments after the story posted, from the principal investigator of the study, the trial does not use a placebo. It is a randomized, controlled study, but the control group is receiving the “standard of care.”

I’ve written about randomized, controlled studies for more than 30 years, and, unless there’s a known effective treatment or the intervention cannot easily be mimicked with a placebo (à la brain surgery), they always have placebo. They also typically blind researchers and the patients from knowing whether they’re in the treatment or control arm. But not this one.

It’s a subtle, but important, distinction, and I kick myself for not asking the researchers conducting it what they were using for the placebo. It’s especially sensitive because the randomized, controlled design for Ebola studies has been deemed unethical by some.

This mistake did not happen because I misunderstood someone I interviewed, mind you. It’s because I thought I understood it better than I did.

Take home lesson: No assumptions.

Carl: I don’t have traumatic memories of putting something spectacularly wrong in print based on a misunderstanding of an interview. Fact-checkers have certainly saved me from lots of blunders. So has a willingness to recognize when I don’t understand something, even after a scientist has explained it several times to me. Of course, I may be defensively forgetting something! What really galls me is discovering that I didn’t interview someone who turned out to have some key pieces of information—or perhaps an entirely different perspective—about a story I’ve written. Usually they email me after they see my story, once it’s too late to take advantage of their knowledge.

Julio: I agree with Jon’s take-home lesson about never making assumptions. More often than not, when writers have made mistakes covering my research or other activities, they have involved making unnecessary assumptions that led them down a rabbit hole. For most scientists, at least good ones, very little is implicit. If we make assumptions, they are clearly stated. I would say this is a fundamental difference between ways of thinking for scientists/engineers/philosophers and those who are not. Challenging assumptions is central to the scientific method, and it comes with years of training, practice, and, yes, many failures rooted in unchallenged assumptions.

So, one lesson here for the writer covering science is to dig a little further and ask the scientist to discuss his or her main assumptions in the study and how they were addressed.

Carl: I’ve encountered another type of dangerous assumption in these interviews: Scientists assume that reporters understand not only the underlying science, but the whole history of the field leading up to their own experiment, as well as the connective logic between the different things that they did in their own research. So when you ask scientists to describe their work, they just leap into a narrative that is missing many of the most important pieces of information for writing an article.

I typically start my interviews by letting scientists tell the story the way they want to without saying much of anything along the way. Then, when they’re done, I say, “There are just one or two things that I still don’t understand.” Then I go all the way back to the beginning of the story and ask questions to fill in all the steps they left out.

Karen: I feel that I have been very lucky in having mostly very-good-to-awesome interactions with reporters. The only disaster I have had was with a student newspaper where they just did not know or fully understand the process of embargoes.

I generally enjoy the process of talking to reporters because when it works, it works really well—a natural conversation of two people exploring a topic, trying to understand what happened and what it all means, and discussing the larger implications. And, in the end I generally know what a future audience will ask, where I need to explain better or differently, where people are not convinced, and where I am confusing.

Geoff: To prevent inaccuracies, scientists sometimes ask to see their quotations in a piece before it’s published. Scientists, do you request this? Journalists, how do you respond to such requests?

Carl: I do sometimes get requests from scientists to see their quotes before my stories are published, and I always explain that it’s simply not possible. No major publication I’ve worked for has looked kindly on such a practice. Journalism only works well if there’s a bright line dividing journalists and their sources. We begin an on-the-record interview with a shared understanding of the ground rules: that anything sources say can be used by the journalists. If the sources get to second-guess the journalists afterwards, the process is corrupted. In some cases it’s true that a scientist might catch an error that shouldn’t get into print. But very often, scientists end up editing quotes for tone, so that they sound the way they’d prefer to sound.

While I don’t show sources their quotes, I do make sure that they are accurate.

Jon: I suspect this debate started at the dawn of journalism.

I agree with Carl, and I draw that bright line in about 95 percent of the interviews I do. But five percent I do not.

Before I explain the exceptions, I want to reinforce what Carl wrote. Sharing quotes ahead of time has many drawbacks. It creates the impression that the journalist is a PR arm of the source, and makes it seem like the product is a collaboration, which is the fiction Janet Malcolm decried in her famous New Yorker essay [“The Journalist and the Murderer”; paywalled].

As for the five percent, I sometimes speak with people who attended, say, a closed-door meeting that discussed proprietary or other sensitive types of information. Maybe they offer to speak with the understanding that I will not quote them. I am not a big fan of blind sourcing, but if I go that route, we have to agree on how they’re identified and, often, what I can and cannot report. I’d rather quote them by name, so I sometimes say, “Look, speak to me freely and I’ll run anything by you that I want to quote, but I want to use your name.” I’d rather have that quote than an anonymous source or no quote at all. When I do this, I explicitly say I do not want them making cosmetic changes, but will consider changes of fact or interpretation. I much prefer to read them these quotes than to send them in writing.

I’m also generous with young people who have never spoken to a journalist, and if they say something that I realize will cause them great harm—à la “My graduate advisor is a moron”—I often then and there will ask, “Are you comfortable with me reporting this?”

Most journalists I know, whether they specialize in science or not, feel comfortable sharing a sentence or a paragraph to have a source check the accuracy. I was surprised to hear David Carr, the New York Times media reporter who recently passed away, tell Terry Gross in 2011 that he did this for another reason: so they’re not surprised by what he writes.

My bottom line is we have to serve the reader, not the source, and while many journalists may not admit to ever sharing quotes or passages, my sense is that many have adopted a policy similar to my own. As a rule, don’t do it. Occasionally break the rule.

Karen: I have never requested to see my quotes, but sometimes I have been able to review quotes or parts of the text for scientific accuracy before printing. I appreciate this courtesy, especially when what I have said or what has been written is not 100 percent accurate.

Julie: I tell scientists that they have every right to ask to see their quotes—and journalists have every right to say no. It’s a conversation.

Early in my career, I regularly asked scientists to review chunks of the text, including quotes. I know that’s anathema for many journalists, but I discovered a lot of errors I wouldn’t have caught on my own, and the problems that journalists imagine—sources wanting to tone down their quotes or shape the story—really never happened.

I’m more judicious about that now, because I’m writing for different publications and writing different types of stories. I’m also better now at fact-checking without relying so heavily on my sources.

However, I’m working on a story now that’s, er, a bit touchy, and many, many sources were hesitant to talk to me about it. In that context, I’ve freely agreed to show people quotes in advance, or even to talk with them and discuss later if I’ll name them or use quotes at all.

Julio: Despite the fact that I’ve been misquoted occasionally, I generally don’t ask to review my quotes or the story. In some cases, where I hear an inexperienced or uncertain voice on the other end, I ask the writer during the interview for a brief synopsis of what we talked about and ask if there was anything that was still unclear in his or her mind. If this discussion still seems somewhat muddled, I’ll actually offer to look over quotes or the text of the story. I find Julie’s approach refreshing and similar in spirit to mine, but from the writer’s perspective.

Daniel: When I ask to see quotes, I explicitly say that I won’t change the substance or meaning of what I’ve said and that the writer is welcome to use my exact quote if they choose to. I have never asked to see a full story or the context surrounding a quote.

If the goal of a piece is to explain the science to a broad audience in a compelling, accurate way, wouldn’t it be advantageous to ensure that the quotes accurately reflect the scientist’s intended meaning? For example, I’ve occasionally used a word in a spoken quote that was ambiguous or imprecise, and permitted a different an inaccurate interpretation. A simple word change fixed it. That word choice made no difference from the journalist’s perspective, but from my perspective as a scientist, it mattered a lot.

This discussion of the “bright line” principle is fascinating to me. I’d guess that it’s not something most scientists have thought much about. From a scientist’s perspective, it’s hard to see the harm to journalists in making sure that quotes are precise and accurate to the speaker’s intent (not just to the exact wording they happened to use when recorded).

Julie: I’m much more relaxed about the issue of showing quotes, I think, than either Jon or Carl is. I explain the ground rules up front, typically saying something like, “I’m sending this along for your help in catching things that are wrong. Let me know if this quote misrepresents your intent, and we can talk about whether some small change will clear things up. But also remember that a big part of the reason we include quotes is to get the sound of a real person talking in a relaxed way, so please don’t tidy your words up into stiff sciencese.” Scientists have caught all kinds of issues I wouldn’t have caught on my own, and I haven’t found that they’ve tried to influence the direction of a piece or make themselves look better or anything along those lines.

At this point, I’m much more likely to read copy over the phone rather than send it along in an email.

Daniel: I’ve done a lot of interviews for journalists who aren’t trying to write pieces about the human experience behind the science. Rather, they are writing a brief overview or summary of a new finding for a short piece in a paper or magazine, and they aren’t delving into the motivations and goals of the researchers. Those are the cases in which sharing a quote seems to be beneficial both for the journalist and the scientist.

I understand that you’re wary of slippery slopes and are upholding a journalistic standard, but would you have as big a concern about journalists sharing quotes with sources for those “coverage” sorts of pieces?

Carl: I can only really comment on my own work and experience. But I will point out that when I was starting out in science writing as a young assistant editor at Discover writing little 300-word stories, I was forbidden to show quotes to my sources. It was a good policy then, and it’s a good policy now.

Jon: This is a complex issue. My sense is that it’s (usually) wrong to share quotes with sources and that Dan is arguing that it’s not a sin if it’s for good things. Dan, to be blunt, I think this sidesteps the issues I raised earlier about the appearance of a collaboration, which borders on PR. It’s like conflict of interest. It’s important to avoid the appearance.

Geoff: Let’s talk about the interview process itself. Scientists, are there any particular types of questions that you wish writers would ask, but that they usually don’t? And writers, are there any particular types of questions that you’ve learned are more likely to yield the information and quotes you need for a good story?

Carl: Some scientists are natural storytellers. An interview with them just feels like hitting the “Play” button, and I can tell I’ll have an embarrassment of riches to choose from when it comes to piecing together the story and bringing their experience and thoughts into it. I just have to make sure we hit all the important steps in the story.

Some scientists are enthusiastic, but mistake me (and, by extension, my audience) for a fellow scientist. In those cases, I have to help them re-orient themselves to a broad audience. Sometimes, once I learn about some particular scientific point, I might say to them, “Is there a metaphor you like when you think about this?” The fact is that scientists love to think in metaphors. They just never get asked to share them.

And some scientists initially feel uncomfortable with speaking as an individual about their science. If I ask how they got started investigating the evolution of turtles or the shape of rivers, they may not know how to begin to answer that question. In those cases, I try to assure scientists that I really am interested in the science, but that I also recognize that science is something that people do.

Daniel: One of the questions I’d like to see journalists ask more often came up earlier in our discussion: What are the assumptions underlying your work and how can you be sure that they’re justified?

Here are more questions that can take a scientist out of PR mode and into a more nuanced and complete discussion of their research.

1) Are there others in your field who would disagree with your claims? (This can be a great way to find out if there are skeptics who would dispute the claim, or those who think it means something different. If the scientist can’t identify anyone who would dispute anything about their claims, they likely aren’t being forthcoming and journalists should tread carefully.)

2) What are the limits to the generality of your results? What else would you need to do to show that the results had wider applicability?

3) What aspect of your results is the most robust? That is, if you had to stake your career on one finding from your paper, which would it be?

4) Which aspect of your results gives you the most pause? That is, if you had to pick one aspect of your results that you think might not hold up, which would it be?

I ask questions 3 and 4 whenever I’m on a Ph.D. committee that falls outside of my specialty area.

Jon: If I’m doing a story about a research finding, I do often ask, as Dan suggests, who might disagree with your results or see them differently. It also helps me if a researcher can point to a graphic they think best displays the main finding.

Carl mentioned a perpetual challenge with some scientists: They speak to an audience of peers rather than to my audience. Even when I’m writing for Science (my main job), the audience has geologists who may not know a lot about cells. There’s a great line in the movie “Philadelphia” when Denzel Washington, a lawyer cross-examining a witness, says, “Speak to me like I’m a five-year-old.” I don’t use that line, but I do sometimes say to biologists, “Explain it to me so a geologist or astronomer could understand you.”

Geoff: Finally, as I often end interviews by asking: Is there anything that I didn’t ask that you think we should have discussed?

Julie: One other thing that occurs to me is that I almost always start by asking about how the researcher got interested in the topic of the study in the first place. I try to begin by getting some grounding in their career as a whole and in the background for the particular study.

I do a lot of mathematics writing, and I find that often, mathematicians simply don’t know what level to speak to me at. Sometimes the problem is that they talk as if I were a fellow mathematician (though not one in their field—they’re savvy enough to get that). But actually, I don’t want them to talk to me like a five-year-old, because they can’t actually say anything at that level, and I can’t get the information I need. So I often tell the mathematician a bit about my background (I have a master’s in math from MIT), and I tell them that I have a fair amount of mathematical sophistication but that they can’t count on me to actually know anything.

Julio: At first contact, when a writer is scheduling an interview with a scientist, it might be good to provide a thumbnail sketch of the “what, where, when” of their story, meaning the length and tenor of the piece, the outlet, and when they expect it to appear, as well as any ground rules for the interview. Also, as a courtesy, please send the scientist a copy or link to your story when it appears.

Sometimes laypeople, students, and colleagues will ask me, “How the heck did you arrive at that particular question, much less the answer?” I find that people are really interested in how science (and the mind) actually works. They intuitively know and are intrigued by the fact that there is a lot of serendipity and accidents of personal history. Oftentimes, this interests them as much as the actual science; it provides parables of not only how science actually works, but also how all of us navigate opportunity in life. In some cases, scientists haven’t really thought through the sequence of events that ultimately led to the particular discovery, so this gives an experienced writer and conversationalist an exciting opportunity to relive the journey.


Geoffrey GillerCourtesy of Geoffrey Giller

Geoffrey Giller

Geoffrey Giller is a TON fellow sponsored by the Burroughs Wellcome Fund. He is a freelance writer and photographer and a recent graduate of the Yale School of Forestry and Environmental Science, where he studied frogs and water contamination. He has written for Scientific American, Audubon, and other publications. A lover of all amphibians, Geoffrey especially adores salamanders and hopes one day to meet a hellbender in the wild. You can follow him on Twitter @GeoffreyGiller and see some of his photography at his website.


  1. Pingback: I’ve Got Your Missing Links Right Here (21 March 2015) – Phenomena: Not Exactly Rocket Science

  2. Thanks for publishing this interesting conversation between some of my writer and scientist heroes. I noticed that the people who drew the clearest line about sharing quotes with scientists were those working for pubs that have the funds to employ fact-checkers. For smaller publications, sometimes the best fact-checker available is the expert interviewed. It’s always a judgment call deciding what’s going to make your writing most accurate in the current climate of downsized editorial teams.

Leave a Reply

Your email address will not be published. Required fields are marked *