Also this week: text mining Rate My Professor, reproducibility is overrated, camera traps vs. Harvard students, overly honest R package help files, and more!
Entomologists are using legos to help work with collections of pinned insects. The pictures in this story are great! (ht: Bug Gwen)
I had never considered the consequences of saying “black” versus “African American”, but this article suggests I should. It can have huge economic consequences, it turns out.
The best thing on the internet this week was written decades ago. Via the latest issue of the ESA Bulletin (which by the way totally kicks butt these days): Thomas Park’s 1960 President’s Address to the ESA. Fascinating to read Park’s (brief) thoughts on where ecology stood and where it was going. Today, many of his remarks seem prescient; interestingly, a few don’t. Some quotes, to give you the flavor:
On doing exciting science and conveying it to others:
I subscribe to the view that every scientist, indeed every creative person, must believe that his efforts are exciting and that this can be conveyed to an audience considerably larger than himself.
On ecology without apology:
I hold the view that the time is past when it is either necessary to tell ourselves what our field is all about or to apologize, either directly or indirectly, for the fact that we are ecologists. We can take pride in the progress of our field and even greater pride in its potential.
In applying analytical methods to ecological problems one does not excommunicate natural history. Rather, one imposes upon it new dimensions of concept and method and thus increases his capacity to interpret events that are otherwise not interpretable through unaided observation.
On field experiments as the next big thing:
Certain ecologists consistently find it rewarding to manipulate the field conditions. But I urge that the trend should accelerate, gain wider adoption, and perfect its techniques.
On the intimate links between ecology and evolution (which were not universally recognized at the time):
To a greater degree we should analyze what [natural] selection actually means in terms of mechanism. We cannot let the word stand, by itself, as if in using it we achieve an explanation. Any selective event involves the ecology of the population experiencing it and is interpretable in respect of the basic parameters of multiplication, survival, and dispersal. It is our responsibility to enter this arena with vigor. The point is not to compete with the population geneticist but to rally about a common problem — a problem to which we have much to offer.
On stochastic modeling (something on which Park didn’t prove prescient):
There is some doubt as to whether probability modelling can attain a pervasive validity in a field so complex as ecology. Population genetics (where the approach harvests rich reward) enjoys a neatly built-in probability mechanism. Population ecology does not.
On the importance of big new ideas:
I come to my final comment which, in a sense, is the one of greatest long-range importance. This is true even though I have nothing constructive to propose. I name the entry: “New ways of looking at familiar things.” The documentation comes from history. Upon the invention of a decidedly better way to marshall evidence, and erect hypotheses, a science so favored increases in vigor and expands its horizons…Let us hope we are approaching this stage in ecology.
At the risk of preemptively undermining a post Meg has in the queue: Simply Statistics argues that reproducible scientific research (i.e. making your raw data and code available) is overrated. It’s far more important to make sure investigators do their studies right in the first place. Click through for a telling analogy with disclosure rules in finance and other industries. Then again, their post doesn’t consider by far the most important audience for your raw data and code: you and your lab members at some future date.
Terry McGlynn reflects on two years of Small Pond Science.
Joan Strassmann says that you shouldn’t acknowledge anyone in your papers unless you get their permission first. Which surprises me (and Stephen Heard), as I’d never heard of anyone who asks for permission or expects to be asked until I read her post. I can see why she’s a little annoyed by the particular example she gives–it’s an oddly-phrased, unclear acknowledgement. And I agree that it’s not appropriate to acknowledge people in such a way as to imply that they hold views they don’t hold. But such cases are very rare in my experience. Worth noting that the authors of the acknowledgement in question emailed Joan and commented on her blog to apologize; they just meant to thank her and it came off wrong. Maybe the take-home message here is just for authors to make clear what they’re acknowledging others for?
I kind of hesitate to link to this because Rate My Professor is a terrible site and I ordinarily ignore it. But I suspect many folks will want to check this out so here goes: this site lets you search Rate My Professor reviews for the frequency of occurrence of any word or short phrase. It shows you results broken down by the gender and field of the prof being reviewed. Turns out that many, many words have gender splits that to my eye look larger than can be explained by pure sampling error or some simple null model. If you do go play with this, I suggest not jumping to conclusions–there are a lot of underlying factors driving this sort of word use data, including a lot of intercorrelated covariates on which you have no data (gender of the reviewer, size of the class, age and experience of the prof…see here for a bit of discussion from the site author). In playing around with this a bit I found that effect size and direction are not always as I would’ve expected. For instance, there are negative words that are more common in reviews of male profs, and positive words that are more common in reviews of female profs. And I would definitely not jump to the conclusion that these data are representative of what students say in teaching evaluations conducted by their universities. Students who use Rate My Professor are a very biased sample of all students, in various ways. (ht Frances Woolley)
How many Harvard students actually attend lectures? The authors of the linked study set up GoPro cameras and did a bunch of image analysis to find out, which seems like overkill in retrospect given that everything they found is pretty much what you’d expect to find. (ht Marginal Revolution)
A tenured prof is about to be fired over a blog post. There’s a lot more to the story, click through for details; I haven’t studied it closely enough to have an informed opinion. (ht Margaret Kosmala)
#overlyhonestRpackagequickstartguide (ht Brad DeLong)
And finally: Knowledge is power, France is bacon. (ht Not Exactly Rocket Science)
Hoisted from the comments: