Friday, May 14, 2010

Cows are apparently sexier than tensors.

As you know, I am the coauthor on a new paper on community structure in multislice networks, which has just appeared in Science and which I genuinely believe makes a very big advance.

I am also the coauthor of a recently submitted paper entitled A Mathematical Model for the Dynamics and Synchronization of Cows that has been posted on the arXiv, as is normal procedure in mathematics and physics (and, increasingly, other fields). Nevertheless, this paper has not yet been peer-reviewed, and I can't stress this enough.

Guess which paper is getting more attention?

I will readily admit that I have some fun when I get media attention, and that it's pretty damned cool to be see articles in places like Technology Review, Boing Boing, and Marginal Revolution about my work. I'm flattered, and some of what is written is entertaining (and some of the accompanying pictures are entertaining). This has been a fun project with some interesting scientific results (on which I believe we can build with a lot of additional hard work), and we included a couple of jokes in the paper because frankly it makes it nicer to read. Also, I am happy to see a lighthearted tone in some of the places that have discussed our work, but I am a bit disturbed by a couple of things:

(1) Some of the venues that have discussed our work used figures without asking for our permission. That's just not cool. I would gladly give permission, but you really ought to ask.

(2) I'm not surprised by the media attention on something that hasn't been peer-reviewed and indeed have experienced this before, but can we please make a distinction between something that is published and something that is on a preprint server and has not yet undergone peer review? Some (but not all) of the articles that I have seen are ignoring this crucial distinction, and it's a bit annoying to see increasingly inaccurate statements from people who are playing a sort of "whispering game" and making incorrect approximations of what other people wrote (and writing unsubstantiated things as if they were facts). I know that's the nature of things, but it's frustrating, and when I see a post that is not lighthearted but instead tries to use our work to advocate some cause that has absolutely nothing to do with what we studied (which is a biologically-inspired mathematical model, and I feel a very interesting one), I feel just a little bit worse about the world than I did before.

If you are reading this and are a journalist, please note that my coauthors and I wish to wait until after the peer review process runs its course (and the paper is accepted to a scientific journal---ideally the one to which we have already submitted it) before we grant any interviews. We are very excited about our work, but please let's not be premature about things! Now if you want to talk to me about multislice networks, then I would be extremely pleased to talk to you now.

Thank you.


Aaron said...

Given my own past experiences, none of this surprises me. A paper in Science is likely to catch a few journalistic eyes simply because it's in Science, but the paper's topic still needs to be catchy. Temporal communities is probably too dry for most of the people science journalists are writing for (not to mention most journalists themselves), so most will look at other articles. Unless, of course, it has a biomedical application, or a "nifty" story angle, or somehow overturns conventional wisdom about social behavior.

Mathematical models of cows, though, that has a good combination of craziness and seriousness that will grab people's attention and make them go "huh. that's interesting." (It's also liable to be listed as example of "government waste in science" by someone like Senator Inhofe.) It's the same reason why mathematical models of zombies got discussed on NPR a while back. If someone modeled how many vampires could live undetected in modern society based on blood production, risk of detection, etc., it'd be all over the news. (Hmm...) Also, I think pat explanations of common social phenomenon register high on the science-journalism-ometer, since everyone loves to think there's a simple causal explanation for why the people around them are screwed up.

Yeah, the part about the "whispering game" is annoying. I've had a little bit of that with my work on terrorism. It was amusing to see both liberals and conservatives mentioning my work to support their contradictory claims about how government should be run. But, I don't know how to deal with that stuff, really, except to try to ignore it or try to explain things very very simply to anyone who contacts me. This is not easy! If it shows up on a blog, sometimes I'll post a small corrective statement to set things straight. (Often people appreciate it when the authors of a study show up in that kind of forum.) I try to be pretty dry and detached when I do that though, since I think that's the best way to respond to dis-information.

Sadly, I think the distinction between published and preprint is lost on most science journalists. To them, a study is a study, and their job is to write about it before the novelty wears off. In that sense, I respect Nature's embargo policy, since it prevents itchy journalists from encouraging bad behavior in scientists (I'm thinking of polywater and other examples of pathological science). It can't prevent bad behavior among journalists, though.

For the honest players, it's a tricky situation, partly I think because journalists (and bloggers) are always on tight deadlines, are almost never themselves very knowledgeable about the topic, and are interested mainly in the story they're trying to tell rather than the nuances, caveats, uncertainties and other "details". Their task is usually to take something complicated and write a short explanation that someone with only a high school science education could understand. It's basically an impossible task. In my experience, the best people in the game, i.e., those most likely to understand what you tell them and those most likely to write a scientifically accurate and interesting story, are the professionals working at high-profile outfits like Nature, Discover, etc. Many of these are former scientists or at least have some advanced training. That's not to say they're perfect (even with the backing of a "fact checking" department), but they're definitely better.

Anyway, congrats on all the coverage and on the pub in Science!

Mason said...

Actually, one of our directions for the multislice stuff can be construed as biomedicine. (We're working with neuroscientists and physicists for this and have some nice results, although we're not ready to write up a paper yet.) It will be interesting to see what happens depending on how spiffy the results ultimately are.

Yeah, the whole cow thing definitely has the 'first makes you laugh and then makes you think' characteristic of past Ig Nobel winning research. I want one. :) Although we were occasionally flippant just because of the subject matter, if we can push this through, this could ultimately help with practical things (e.g., helping in design of pens for cows). We are admittedly a long while away from that stage of the game.

The whispering game must have been horrible with your terrorism research. I decided not to post any comments on the animal activism blogs where I saw posts. I want to steer clear of that, and Oxford has a bit of an "interesting" history with animal activism anyway (via massive protesting over new experimental facilities), so I don't want to be involved in discussions that are so far from my expertise. However, I'd be delighted to tell people why things like the Poincaré-Bendixson theorem need to be generalized for piecewise-smooth dynamical systems. I commented on one of the lighthearted posts, but that was in large part because I went to school with the author of that post and also because he specifically posted it as an arXiv entry (and knows the difference between something that is peer-reviewed and something that isn't!).

The science writers for venues like Nature in general do seem quite good. For what it's worth, I have had very good experiences on occasion---e.g., the Wired writer who wrote about my baseball networks research (while it was a preprint, actually, but in that case I didn't need to worry about animal activists!).

My website has had something like twice has many hits as usual today (granted, that's still a small number), so we'll see how long this continues. I still have a bit of time to deal with things for a couple of days before the next work spike arrives...

Justin said...

The norm in astronomy is that a paper is not posted until it has been accepted, which avoids a lot of the problems you mention. To me it seems bizarre and presumptuous that anyone would post something that has merely been submitted...

Mason said...

Justin: Posting preprints is the standard. It actually helps science's self-corrective process substantially to have things posted early, because then with feedback the journal version of the paper can be even better. Occasionally, something like this happens too.

Michael F. Martin said...

I have no idea what "whispering game" you're talking about -- it sounds ugly, sorry -- but I don't see how it should ruin yours and your coauthors enjoyment of some well-deserved recognition. Do most of the people who read the article care about the details that will be ironed out during peer review? No.

I'm a non-specialist who found the article interesting because I simply haven't seen any similar models of synchronization dynamics in animal herds. I'd like to see a lot more of this kind of work, and I hope the attention you're getting will encourage you to continue, and others to join you.

Like I said, I have no idea what "whispering game" you're talking about, but it sure sounds like plain old envy, of which I have seen a lot.

Michael F. Martin said...

Oh, and I would love to read the Science article you'd prefer for me to read, but it's gated, and my law firm does not have a subscription.

Mason said...


Thanks for the comments! And I'm glad you like the article!

I do believe the cow article is very much worth reading. I very much enjoyed working on that project, and I am looking forward to building on the work described therein. I am not unhappy with the attention, but I get disturbed when comments of "this might suggest that [blah]" turn into things like "this work shows that [blah] and therefore we should [insert cause of person]". It's a math paper---and in my opinion an interesting one---and I am very pleased for people (both specialists and non-specialists) to enjoy it both before and after pee-review. But there is a difference between the two, and I think people should care about that---not to ruin enjoyment or not to read it (otherwise, I wouldn't have posted it on a preprint server before it was suitably revised and accepted) but to appreciate the fact that there is a strong sense in which the work is in-progress that makes it different from a published work that has already passed the muster of peer review.

Some of the comments in those blogs had a bit of a peanut gallery, and that might well be the jealousy you're mentioning---but I think the whispering game is more a matter of some of the writers not worrying about changing important meanings of non-technical items through exaggeration. (Obviously, the technical aspects of the science need to be made more digestible...)

I can give you a link to the accepted version of the Science article that will not require you to have a subscription: You can fine that here. If you e-mail me, I am also happy to e-mail the version exactly as in Science, but I am not allowed to post it on my website.

Also, it's not a preference for what people read---it's just that it's crucial to treat things that are preliminary in a strong sense different from things that have already undergone an important (and hopefully stringent) process.

Michael F. Martin said...

If you weren't disturbed by sloppy reasoning, then I suppose you wouldn't be a good mathematician. On the other hand, we have the wise advice not to let perfect be the enemy of done.

Perhaps more importantly, you're going to have to gloss over some details in making your work more accessible to non-specialists. This will inevitably expose you to criticism from specialists -- especially the jealous sorts.

What motivated me to forward your article to Tyler Cowen -- actually, I forwarded him the arXiv blog summary of your article -- is my desire to see more network dynamics , especially synchronization, in economics. We're in the midst of a nasty financial mess; nobody understands fully what happened; it sure would be nice if some smart theorists could explain some of these anomalies:

Anyway, thanks again for the interesting papers and I sincerely wish you the best of luck in making progress.

Mason said...

For what it's worth, Tyler's blog entry is not one of the ones that bothered me. It's the animal rights ones that caused me to take a step back. (On my end, it's a bit disturbing for my stuff to be used for somebody's righteous cause using conclusions from our paper that we don't really have---and that is independent of what I might actually think of the cause.)

My stuff is certainly far from perfect, but yes I do tend to anally obsess about things sometimes. :) Anyway, things are only really done when you get bored and move on to something else.

I am fine with glossing over some details. I was not fine with statements about the work possibly suggesting things being phrased in some places as if it were irrevocable proof of something. That's very different. I in fact enjoy writing expository things about science, and you're absolutely right that technical details have to be glossed over.

My collaborators and I have actually done a small bit of work recently on dynamic networks in finance---not with synchronization per se, however. One thing one can see from exchange rate data is that "network roles" of certain currencies exchange rates changed pretty quickly amidst the camel's back breaking. We're pretty far from having an explanation, and to be honest from my perspective, it was an interesting example of time-dependent networks. (My coauthors from HSBC obviously had a different motivation.)

Thanks for the link---I actually vaguely remember seeing that one before.

I haven't seen criticism about the cow work from specialists yet, but we'll see what my referee reports say! :P

Michael F. Martin said...

The animal rights blogs? I didn't see those. In that context, your remarks make much more sense.

As a lawyer, I would have been sure to advise you to include the obligatory disclaimer: "No cows were actually blown up into spheres during the course of these experiments."

That's interesting about currency exchange rates. I'd love to read more about that. I'm sure your HSBC collaborators are equally eager for me not to!

Mason said...

Yeah, I was happy to be on the lighthearted, geek, and science blogs! The other stuff can be a bit scary, though.

We have two papers so far on financial networks (with more on the way, though not for a little while). One of them has been published and the other is a sequel to that paper that is currently going through peer review.

published one

not yet published one (sequel to the first paper)

In fact, one thing on the docket is to use the new method we developed in the Science paper that is better for time-dependent networks than the "old" available methods (which weren't actually all that old).