What it's like when Questlove does a better job tweeting about your research than CNN

By Holly Buck and Gernot Wagner

A conversation between Holly Buck and Gernot Wagner

Questlove tweet

Holly Buck: So how has it been with CNN, Questlove, and a parody account for Chance the Rapper all tweeting about your co-authored paper in Environmental Research Letters?

Gernot Wagner: Suffice it to say, it’s been an odd experience these past three, four days or so. A couple weeks ago, when we were asked about a release data for this paper, my very first thought was that releasing this on Friday after Thanksgiving wouldn’t be a bad idea. The only people who pay attention are the ones who typically pay attention to this stuff.

As usual, there will always be some random blogger somewhere who picks up a solar geoengineering paper – it’s typically someone for whom this is the first time hearing about solar engineering – and then proceeds to spin it completely out of proportion.

The difference here is that this time it was CNN who committed the journalistic malpractice here. That’s the story that proceeded to go viral. It was trending as the number one breaking news topic on Twitter a few hours on Friday night. The unfortunate coincidence there was that the US Government Climate Assessment came out on Friday as well. Meanwhile, the CNN piece wasn’t even written by a journalist as far as I can tell, but by someone whose bio says “Trending News Producer.” He was basically looking for some sexy climate story, and he somehow found this paper.

CNN wasn’t even the first to write about it. The Guardian was, and they actually wrote a decent story. Damian Carrington pinged me to talk about the story. [Co-author] Wake [Smith] and I spoke to him the day before Thanksgiving, he got some other comments, and it turned into a good story that went up on Friday morning. It was a very sensible piece about how this new research paper – narrow focus, narrow answer that itself wasn’t all that surprising, but that does provide a sensible step forward in these discussions. So the Guardian story puts it all in that context.

And yeah, that Guardian story got about the right amount of attention, I’d say. A few dozen retweets, a dozen or so comments. Perhaps a bit more than usual. There, too, of course, is always somebody who takes things out of context and spins it to his or her liking. But oh well.

But then there was the CNN story. I’m looking at the article’s Altmetrics page right now. It’s just completely out of whack. The article came out Friday. By now it has over 30,000 downloads. It’s in the 99th percentile on any of these metrics, which is completely insane.

Holly: That in itself is really interesting.

Gernot: The CNN story is bad, and, of course, given that CNN’s feed is apparently fairly popular to copy among news outlets, it doesn’t take much for that to go viral. Let me put it this way, there are apparently plenty of lesser journalistic outfits, who basically simply appear to copy popular pieces from elsewhere to get in on the action. They barely appear to paraphrase the CNN story, don’t credit CNN by the way, but use the exact same quotes. So now there’s all these other places – from The Hindu to Metro to the New York Post to Breitbart – that have a version of the story up. They were clearly all just in it for the clicks while it was viral.

Holly: It’s like the first person to set the frame gets to lay the narrative.

Gernot: Yes and no. The Guardian was first, but they basically had the appropriately boring framing – not quite “researcher publishes research,” but close: “Solar geoengineering could be ‘remarkably inexpensive.’” That itself is a framing that sometimes goes viral for the wrong reasons, but that’s basically an accurate description of what we did. The CNN framing – “Harvard and Yale scientists propose solution to climate change” – of course, was entirely wrong. Let me not even get into the affiliation question here, and whether an economist is a “scientist,” but, of course, the bigger deal is threefold: that CNN presented this as novel, as a “solution”, and said that we propose to do it. Just to be clear, solar geoengineering isn’t novel, it isn’t a solution, and the paper says nothing about whether the world should do it.

Holly: Is Trump right when he says that CNN is incompetent? Like, does he have a point that our media system is broken?

Gernot: Let me give you a pointed answer: Yes. But this is a very different quibble here, than frankly, Trump has, right? And I mean, not to feel too much pity for the guy, but I’m sure there is a lot of this type of stuff happening to him. Just to give you a sense, and don’t ask me how I know this, but this story was competing for awhile for being the number one trending news story on Saturday, with some story about one of the Kardashian sisters spending Thanksgiving somewhere. I don’t know, in Ohio or somewhere. Doesn’t really matter.

Anyway, in that sense, yes, Trump is absolutely right, that places like CNN nowadays make it easy, because frankly, there’s so many news outlets out there, one of them is going to get it wrong, and chance is pretty good that the sexy sounding wrong message, or the one that was optimized to sound sexy, is the one that goes viral and swamps everything else. Journalists are now competing with “content producers,” who put up a bazillion of these stories on any given day. They hope for one of them to go viral which drives eyeballs, drives ad revenue. Twitter is happy about this. CNN is happy about it. These other 50 outfits are happy about it, that eyeballs are being drawn to their website.

Holly: So, what does that mean for researchers, knowing that we have this toxic swamp of a media ecology? I mean, is it incumbent upon us to change our behavior because we know that’s out there?

Gernot: I guess really what you are asking is, “Should this paper have been written?”

Well, my short answer is, “Yes, it should.” Of course, I’m biased here. I’m one of the co-authors. But yes, as far as I can tell, it’s one of only three “modern” costing studies of SAI. There’s one by McClellan et al., there’s one by Robock et al. And then there’s a bunch of reviews, and, frankly, lots of urban lore on this topic. For example, there’s this perception that a dozen modified business jets can do this. Basically, a rich guy takes the seats out of his Learjet and starts modifying the global climate. Well, no.

So yes, this study, I’d like to think, adds an important bit to that conversation: While we confirm that it’s technically doable, and that it’s cheap, it would take a new plane design to do it. That’s basically what Wake Smith found in his initial explorations. By the way, he’s the ex-CEO of an airplane modification company. So yes, you could have imagined the at least perceived conflict of interest had he found that modifying a plane could do it. Turns out, it can’t. That’s essentially when we joined forces and said, let’s do this. Let’s do the proper analysis. Let’s write it up. There’s something there.

Now, should these sorts of interactions among researchers be changed because of the potential for their work being taken out of context eventually? No. Again, I’m biased here, of course, but no, the research itself can’t possibly be influenced by that.

Now, should we consider all of this when publishing a paper, how to talk about it, how to describe it on your own Twitter feed, or if and how to interact with journalists? Of course, yes.

Holly: I think that with this particular paper, unless you had re-focused it pretty drastically, it would be hard to message it differently to avoid this kind of outcome. If you were only dealing with journalists that you knew would do a good job, then you could convey, “These are the key points. This is the context,” but since that infrastructure has deteriorated, I don’t know how you get around it without writing a totally different paper.

It seems to me that we already knew that the operational costs are going to be pretty small. And if you think of the whole intervention — you would want to be removing carbon dioxide from the atmosphere probably, which is going to be really expensive. And that would be the overwhelming amount of the overall program cost. Would it be impossible to write a paper that looks at all the costs for an SRM-CDR program? To write a paper with a different context?

Gernot: Hm, you say that we have known that solar geoengineering is cheap. I would question that premise, basically pointing to the premise of the paper: yes, plenty of us have been talking about how everything we know about climate policy reverses with solar geoengineering. It’s “cheap, fast, and imperfect.” It’s “free driver” instead of “free rider,” etc. Guilty as charged. But some of that – lots of it – was what I might call rather flippant, or more urban lore than research. Witness the bits around modified business jets, which, we show here, can’t, in fact, fly to 20 kilometers.

Now, could we have done this completely differently, and framed it as part of a broader solar geoengineering costing exercise, one even including CDR? Well, my short answer would be no, not really. You know, of course, SRM and CDR are completely different.

There’s clearly work to be done on both. On the CDR side, the National Academies just came out a couple weeks ago with a comprehensive review. Very different type of work. I’m certainly no expert on any of that.

Of course, any result then needs to be properly framed. But in some sense, it’s very difficult to do that in any way other than putting a couple sentences on top of a paper, a couple sentences in the conclusion. Could the term “direct engineering costs” have shown up in the title? Sure, the title was our choice. Instead of saying “SAI Tactics and Costs” we could have called it “Narrow SAI Engineering Costs” or some such. There might be a way to add more adjectives here to make sure that no one takes it out of context.

Holly: One opportunity might have been to lay out some further research, so you could list all the things that would be other related costs that we don’t know about, and what scope of effort would be required to answer those questions. Because I’m curious about that.

Gernot: Yeah.  Actually, yes, that’s a fair criticism. What papers typically end on is a section that says, “and here is all the further work required.” It’d say it’s ethically somewhat dubious when you basically stack prior papers with short lines of how the most important next line of inquiry is X, while you already are working on X yourself. So, in some sense I personally always resist that, but actually I can see where this could be used to put the work itself into context.  Though here, with this trending news producer at CNN, it certainly doesn’t look like he has read this piece at all.  So having six pages of caveats at the end of a ten- page piece isn’t going to change anything in a situation like this. It’s just not.

Holly: In retrospect, after this Twitter fiasco, is there anything you would’ve done differently?

Gernot: I think that the short answer is it may be too soon to tell. First, I’m not sure I’d know what to do differently. One option could have been to basically go all out and essentially launch some sort of publicity blitz ourselves. Start a twitter thread, create a video abstract, write an op-ed about it, write blog posts, personally email journalists and say, “Hey, this piece is coming out next week, are you interested?” I avoided this here precisely because I was afraid someone might take it out of context. That clearly didn’t go so well.

Holly: Anything you’ll do different next time, the way you present things? Would you change your research based on this?

Gernot: Let me ask you what would you do in this situation? Would you change your research based on this?

Holly: Maybe I would’ve made an infographic, like one graphic that shows kind of all the pieces of an eventual cost and colors this one line. And then that could be tweeted, and we could be like, hey, we know the answer for this one line, here’s the other things that we would have to fill in with interdisciplinary research in the next three decades or whatever. I think that’d be a cool way to contextualize it.

Napkin Diagram

Gernot: Yeah. I like that idea. I mean, to make sure that’s sort of the most viral thing out there, right? The most retweetable thing out there is in fact something that puts it in context. One of our deliberate decisions early on was not to have another graphic. In the beginning we thought, hey, wouldn’t it be cool to have a drawing of the plane. Here we describe what the plane might look like, but the paper doesn’t actually show what it could look like. That would have been horrible, though, right?

Holly: Yeah.

Gernot: So yes, we nixed that idea.

Holly: Good call.