Friday links: zombie ideas in evolution and psychology, a world without statistics, and more

Also this week: “charging a cover” for students to attend lab, the (ridiculous) ecology of Tatooine, and Florence Nightingale vs. Twitter trolls. Oh, and I make fun of some pseudoscience.

From Jeremy:

Jerry Coyne argues that there are no ring species. I don’t work on speciation, so I’m not qualified to judge this. But as a connoisseur of zombie ideas, I’m curious to hear from folks who do. Is Coyne right? Are ring species now a zombie idea? Are people attempting to save the idea by broadening the notion of what constitutes a “ring species” (as at least one commenter on Coyne’s post tries to do)? Curious to hear what folks more knowledgeable than me think of this. (ht Ed Yong)

Speaking of zombie ideas, here’s one from psychology: the famous Stanford Prison Experiment. You’ve probably heard of it–but you probably weren’t aware that the results actually weren’t clear cut, that the results may well have been driven by leading instructions to the participants from the lead investigator, or that a major replication attempt produced dramatically different results. But then, you’re probably not a psychologist, so it’s not particularly surprising or worrying if you aren’t aware of all the really serious criticisms of this experiment (I wasn’t). What is worrying is that psychology textbooks that cover the Stanford Prison Experiment (which most do) mostly either present the claimed results uncritically, or else present only cursory coverage of ethical criticisms while skipping over substantive criticisms (see the link for details). Textbook authors are quite rightly drawn to classic studies and dramatic examples. But are you really doing students a service if you ignore serious criticisms of classic studies, and focus on dramatic examples that may well be unrepresentative flukes? Plus, as the linked post points out, the controversy over the Stanford Prison Experiment is a great opportunity to teach students about research methods, pitfalls in study design, the importance of replication and alternative explanations, and more (an opportunity that a few textbooks do take up, to their credit). All of which gives me an excuse to link to our old post on whether and how to teach scientific controversies. (ht @TimHarford)

What would the world be like without statistics? Andrew Gelman speculates that it might not be all that much different, or all that much worse.

This week in pseudoscience: Did you know that, because the Concorde deviated from the allometric relationships that apply to other aircraft, it was an “evolutionary dead end”, obviously suboptimal and therefore doomed to fail? That’s what some engineers are claiming, anyway. This seems like a major methodological advance, being able to identify things that are doomed to fail just by looking at whether they deviate from allometries. Wonder what you’d find if you applied that reasoning to vertebrate…OH NO WE’RE DOOMED!!!11!:

(image modified from here.) Also, I love that their own figure actually shows that all airplanes deviate systematically and substantially from the allometries that govern animal locomotion. Let’s not even get started on how the Concorde actually did fly, and failed because of economics rather than the laws of physics. As for the lead author’s notion of a “constructal law” (all systems–living organisms, airplanes, river networks, economies, you name it–must maximize their “flow” if they are to survive)…how, um, interesting. Surprising that a reputable journal would publish this, though perhaps they published it for the data on airplane allometries and decided to pass over the odd bits. Thanks-for-nothing ht to Marginal Revolution.

This is old, but it’s quite interesting (and I bet I’ll surprise a few of you by recommending it): in praise of theories based on verbal stories rather than mathematical models, and empirical studies emphasizing description of variation, “stylized facts”, and what ecologists would call “natural experiments” over sophisticated statistical estimation or hypothesis testing. It’s from macroeconomics, but it’s pretty accessible; you should be able to at least get the gist. Definitely worth thinking about how it might apply to ecology (the author spends a lot of time comparing economics to the natural sciences). I don’t know that I agree with it in the context of ecology, but it’s thought-provoking. Maybe I’ll try to find a way to work it into my Ignite talk for the ESA meeting… (ht Brad DeLong)

Semi-relatedly: teaching students about existing theories vs. teaching them how to develop new theories. Or, teaching them how to solve math problems vs. teaching them how to develop math problems that are worth solving.

BioDiverse Perspectives poses four questions for biodiversity science, and invites you to comment or tweet your answers. The questions are pretty broad and allow a lot of scope for interpretation, so not surprisingly the answers so far are all over the map. But still, it’s good to step back and think about this sort of big picture stuff periodically. We have several old posts relevant to the first two questions, which have to do with translating science into conservation policy (here, here, and here).

Would you “charge a cover” to undergraduate students to attend lab? That is, assign prelab quizzes that students can only pass if they’ve familiarized themselves with the assigned background material. Any student who doesn’t get 75% on the quiz is not allowed to participate in the lab (!) My first reaction is that it sounds like a good idea in principle, but that in practice it would be hard to design and explain it in a way that doesn’t piss off lots of students. I’d first try to come up with other ways of ensuring that unprepared students don’t waste the time of the TAs and the prepared students (not that that’s easy…)

The ecology of Tatooine makes no sense. 🙂

And finally: what if great scientific discoveries of the past had been announced on Twitter? The double helix and evolution are the funniest. And Florence Nightingale shows how to deal with trolls. 🙂 (ht Scholarly Kitchen)