Big Data or Pig Data? Via Deborah Mayo, a great fable for our times from computer scientist Remeez Rahman. A reminder that data doesn’t interpret itself, no matter how much of it you have. Also a cautionary tale of the dangers of putting too much emphasis on prediction, at the expense of explanation and understanding (or even as a way of demonstrating understanding).
And since we’re talking about George Orwell (Rahman’s fable was inspired by Orwell’s Animal Farm), here’s Paul Krugman drawing on Orwell to explain why sharp language like “zombie ideas” has a place in serious, substantive debates. Unlike Krugman, of course, I’m involved in scientific rather than political debates; he’s engaged in a different mode of discourse than I am. But for reasons I’ve explained in an old post, I do think that even in science there’s sometimes good reason to use sharp language. It’s a way to get the reader’s attention and break through complacency–such as the complacency quite naturally associated with belief in widely-believed, rarely-questioned ideas.
Special issue of Nature this week on women in science. I found this review of several recent biographies of female scientists to be especially thought-provoking. In it, Patricia Fara argues that even modern biographies engage in subtle stereotyping, do female scientists a disservice by refusing to criticize them, and over-emphasize the uniqueness of female scientists as individuals, thereby perpetuating the stereotype that you have to be “weird” to succeed as a woman in science. I think she’s on firm ground with the first two claims, but I don’t entirely buy that last claim. After all, don’t biographies of male scientists also often play up how weird they are? Indeed, shouldn’t we expect that scientists who are successful enough to be worth writing biographies about often really will be weird in some ways (while also being perfectly ordinary in other ways)?
The new NSF Division of Environmental Biology blog discusses the sequester and other DEB-related topics, with links to related discussions in the blogosphere. They also seem to be subtly disappointed that they themselves haven’t been deluged with comments yet, based on how much active discussion is going on elsewhere in the blogosphere. To which I can only say: give it time! Speaking from personal experience, no blog builds a readership, much less a commenting community, instantly!
A while back, I asked if scientific misconduct is especially rare in ecology and evolution, or if it just looks that way because misconduct is harder to detect in ecology and evolution than in other fields. In the comments, the strange case of famous evolutionary biologist Robert Trivers was raised. Trivers recently wrote a short book accusing one of his own collaborators of fraud on a 2005 Nature paper they co-authored. He wrote the book after trying and failing to get Nature to retract the paper and publish a detailed analysis of the fraud. Via Mousetrap, I’ve just learned that the book is now available here for free as a pdf. I just went and skimmed it, and I think Trivers makes an overwhelming case. It’s a scary read, because what happened to Trivers could happen to anyone. He’s hard on himself in retrospect for allowing his collaborator the opportunity to commit fraud. But really, he didn’t operate any differently than most of us (including me) have operated when engaged in collaborative work. So forget monsters under the bed or killers hiding in the closet–this book will make you afraid of what your collaborators might be doing with data collection and analyses that you have no easy way to double-check!
Easily Distracted has a nice addition to the ongoing debate over whether massive open online courses (MOOCs) are going to totally change higher education or become a crucial complement to it or have no effect on it or what. He’s a skeptic of the MOOC hype, but also sees some positives, such as that MOOCs will be valuable mostly for positive externalities (like killing off older for-profit online education outfits), and for getting more profs engaged with the broader public. (HT Brad DeLong)
The perils of perfection: Evgeny Morozov on how modern tech companies are mostly producing solutions in search of problems, in the misguided view that life can be perfected–or that we’d want it to be! I do think some of this is relevant to ecology–the most passionate evangelists for “open science” and “big data” arguably are guilty of at least mild versions of the sins Morozov identifies–but I mostly just wanted to throw it out there because I thought it was thought-provoking. Morozov’s new book on this topic is reviewed in Nature this week. (HT Felix Salmon)
And finally, connoisseurs of terrible graphs will appreciate this. 🙂 (HT Felix Salmon)