Friday links: highly significant increase in marginal significance, hidden female authors, the evolution of Groot, and more

Also this week: smartphone microscopes, ranking universities way back when, grade inflation and what to do about it, trading sex for co-authorship, why it feels like you work 80 hours per week even though you don’t, how to write fast, world’s oldest path diagram, academic urban legends, shark vs. shark-cam…[deep breath]…and more. The internet was on fire this week! Oh, and how to defend your thesis. Using a broadsword. ๐Ÿ™‚

From Jeremy:

Ranking US colleges and universities in 1911, and now. The more things change, the more they stay the same, at least for private universities. Some fascinating history here, of which I was totally unaware (click through to the post that the linked post discusses).

Here’s something that has changed about US universities since 1911 (well, ok, 1940), and in a big way: the grades they give. “A” is now the most common grade (!), and almost 80% of grades are A’s or B’s. The linked post discusses small studies of two policies to combat grade inflation: obliging profs to grade on a curve, and providing the average mark in the class on the student’s transcript. I was also interested to see data confirming the stereotype that grades run lower in the hard sciences than in other fields. (Interested because it contrasts with my own personal experience. When I was an undergrad at Williams, the average grades in sciences, social sciences, and humanities there were almost exactly the same.) And I was depressed but unsurprised to see that, when you force high-grading departments to lower their grades, students stop majoring in those fields and give those profs harsher teaching evaluations. Conversely, when you start publicizing the higher-grading departments on student transcripts, more students major in those departments.

Andrew Gelman on “the scientific surprise two-step“: researchers defending their results from statistical criticism often emphasize that their results were expected based on well-established scientific theory (as opposed to far-fetched results discoverable only via p-hacking). But when pitching their results to selective journals and their reviewers and readers, those same authors often emphasize how surprising and novel their results are. Discuss.

The percentage of papers reporting marginally significant p values in the abstract has increased dramatically since 1990. So has the percentage of papers reporting marginally non-significant p values in the abstract, though the increase is much smaller. The results are mostly due to changes in biology abstracts, rather than to physical science or social science abstracts. Argue amongst yourselves whether that says something good or bad about biology. (ht Retraction Watch)

How sloppy citation practices propagate academic urban legends. Citation practices aren’t the only problem, of course–there are plenty of academic myths that don’t arise from sloppy citations.

Are female authors of scientific papers more likely than male authors to hide their gender by using their first initial? This analysis suggests that the answer is yes, though the estimated gap in first initial usage probability is small. The analysis depends on a mixture model and I don’t have a good sense for how well that approach works here, but as far as I can tell it seems like it works fairly well.

The EEB and Flow’s guide to preparing for, and surviving, the ESA meeting.

The 2013 ISI impact factors for the top 40 ecology journals. I’m kind of morbidly curious if anyone will slam me merely for linking to this (Go right ahead! It’s been a long time since someone’s ripped me in the comments, I kind of miss it.) I find these data most interesting because of the changes over time, and what those changes suggest about our collective reading and publishing habits. I’m old enough to remember when Ecology Letters and Methods in Ecology and Evolution didn’t exist and the biogeography journals were well below journals like Ecology and Am Nat (weren’t they?) rather than well above them. Presumably the rise of the biogeography journals is at least in part because of the increased global change focus of ecology? Also worth noting that lumping all these journals together as “ecology” journals runs roughshod over a lot of variation in their audiences and goals. Some of these journals are in totally different fields than others.

An anonymous survey of 400 European economists finds that 94% have engaged in at least one dubious research practice. The more commonly-admitted practices include refraining from citing work that contradicts your own (admitted by 20%) and copying your own previous work without citing it (admitted by 25%). 7% admit to using tricks to tweak the outcomes of statistical tests. And 1-2% admit requesting or offering sex in exchange for co-authorship or promotion (! man I hope that’s an overestimate due to sampling error…) Perceived pressure to publish was positively associated with the admission of several dubious practices. (ht Marginal Revolution)

Arjun Raj asks a good question: since academics don’t actually work 80 hour weeks (or anything close), how come they feel like they do? This seems like a good example of how myths and urban legends arise–it’s people believing stuff that just feels right to them.

Another from Raj: how to write fast.

Here are the path diagrams from Sewell Wright’s original (1921) path analysis paper. (ht David Giles)

Myth-busting: the evidence that making your paper open access causes it to be cited more often is really weak. I’m with Phil Davis on this one: if this is a question we sincerely care about answering, someone should figure out how to do a proper randomized controlled trial (well, another one; the one that’s been done finds no effect of open access). Because just collecting more observational, correlational evidence is not going to shed any light on the matter.

From what plant (or fungus) did Groot evolve? Much as I love the suggestion that he evolved from kudzu, my money’s on some kind of human-plant hybridization event. Because, dude, species from different phyla totally can hybridize, thereby instantly creating whole new taxa without the need for any of that Darwinian evolution crap. ๐Ÿ™‚

We’re gonna need a bigger boat shark-cam. ๐Ÿ™‚

And finally, xkcd has good advice for your thesis defense. Presumably the snake fight portion. ๐Ÿ™‚

From Meg:

This article talks about the value of writing in 10 minute chunks of time. This relates to something that I included in my post on navigating the tenure track. Doing a little bit of writing every day is advice I was given as a grad student, and is advice Iโ€™ve passed on to people in my lab. Though, as Ellen Simms pointed out on twitter, perhaps the biggest challenge is finding ways to find data analysis in โ€“ that is very hard to get done in 10 minute chunks. (Related: my scramble to get an ESA talk together this week, resulting in me only a few Friday links!)

The Guardian has a piece on how an all-male panel is no way to honor Rachel Carson. I agree!

Scicurious tweeted about this microscope attachment for smartphones that has 30x magnification. Sounds pretty neat!

9 thoughts on “Friday links: highly significant increase in marginal significance, hidden female authors, the evolution of Groot, and more

  1. Regarding that smartphone microscope, I kicked in money to this Kickstarter project (The MicrobeScope) that claims to get very powerful magnification with a SmartPhone. It is supposed to arrive next month and I’m curious to see if it really works!

  2. How a myth debunker becomes a debunking mystic:
    (Sorry for highjacking this thread, just documenting how following up one of your links lead to other -er- interesting surf-rides.)

    Following your link to behind “How sloppy citation practices propagate academic urban legends” there was some more ferreting to be done until I reached Dr. Mike Sutton as a major debunker of the decimal point error myth. I even went on to check Emil Wolff’s 1871 publications of nutrition tables, because he was often incriminated as the perpetrator of the error. Whatever Wolff has done, his tables do not present spinach as outstandingly rich in iron. Other vegetables in his tables have much higher contents (e.g., lettuce, rutabaga). So Sutton seems to be correct in his debunking, here. But recently, he embarked on a wild goose chase trying to convince everybody that Darwin and Wallace did indeed read and plagiarize Patrick Matthew’s book On Naval Timber (he recently published a book on it called Nullius in verba). If I remember correctly, Matthew (1831) explained his evolutionary ideas in an appendix to the introduction of his book making bold claims about the importance of navigation for trade and culture, while the rest of the book is about arboriculture and ship building. Anyway, what I gathered from the media hype about this ‘discovery’ is that Sutton used big data analysis to show that some scholars in Darwin’s and Wallace’s social vicinity have read and cited Matthew. But have they read the appendix and mentioned it to Wallace or Darwin? And even if knowledge contamination did occur, how is that plagiarism?

    • Meant to say that the introduction of Matthew (1831) makes the bold claims about navigation and culture, whereas the appendix makes the bold claims about natural selection.

      • Just took another look at Matthew (1831) at While it is true that there are more passages in the book that sound evolutionary, many of them occur in Part IV of the book called “Notice of Authors who treat of Arboriculture.” Alas, while it is already difficult to digest Matthew (1831), it will probably be an awful mess to sort out what passage in part IV is his original thought and what were the ideas of the other scholars on arboriculture reviewed by Matthew.

        Ugh, the though of having to work that out overwhelms me, I guess I put the onus on Sutton and send him back doing his homework (read the sources reviewed by Matthew that contain the various evolutionary sounding passages) before serving up a fraud myth.

        Otherwise, I’ll just conclude that natural selection seems to have been a widespread idea (in the air) and knowledge contamination is not plagiarism.

Leave a Comment

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.