Friday links: statistics vs. the humanities, ecology vs. conservation, and more

Also this week: big new dataset on gender bias in scientific publishing and peer review, Ohio vs. evolution, ecology bloggers wanted, comedy wildlife photography, and more.

From Jeremy:

The Royal Society of Chemistry just released a report on a massive, detailed study of gender bias at its journals. Here is commentary from Athene Donald. The most notable result for me was that there is a detectable association between author gender and acceptance rates, but the effect is small enough to need a massive sample size to detect with high confidence. For the equivalent data for British Ecological Society journals and a couple of other EEB journals (the results for which mirror those for RSC journals in some ways, but differ in other ways), see the work of Charles Fox and colleagues. One thing I like about Chuck Fox’s work is that it combines massive sample sizes and formal statistical analyses to good effect. For ESA journals see Whelan and Schimel (2019).

Writing in Conservation Biology, Hintzen et al. (2019) use machine reading (aka text mining) of 32,000 papers in 16 ecology and conservation journals to show that the fields of ecology and conservation grew apart from 2000-2014, as conservation papers increasingly focused on social and political aspects of conservation. ht Manu Saunders, who comments thoughtfully, and critically. Manu feels that Hintzen et al. didn’t choose journals representative of each discipline, and that their interpretations of their results go beyond what their data will support. Re: the chosen journals not being representative of each discipline, here’s my question: how would you choose a set of journals representative of–or even constitutive of!–“ecology” or “conservation”? Maybe rather than trying to answer that (unanswerable?) question, it’s best to avoid it. Instead, just document how the content of various journals has changed over time, relative to the content of other journals, and then try to explain those changes. Indeed, that’s my preferred way to interpret Hintzen et al.–as showing that the content of various conservation journals and various ecology journals have moved apart over time. Now, would you get a different answer if you chose different journals? Maybe, or maybe not. But either way, changes over time in the content of the chosen journals are still data that seem to me to be worth explaining. And would you get a different answer if you looked instead at changes over time in the mix(es) of journals in which individual authors publish, as Manu suggests? Again, maybe, maybe not. I agree that would also be interesting to look at, but it doesn’t mean that looking at changes over time in journal content is the wrong thing to look at. Anyway, my own gut instinct is that Hintzen et al. have detected a real trend, that would also be detectable in various other ways. But my gut has been wrong before! I also feel like conservation biology started out as a very ecologically-focused field. So it’s almost inevitable that over time it would drift away from ecology “sensu stricto” (whatever that is…) as it came to incorporate more sociological and political considerations. After all, sociological and political factors undoubtedly have massive effects on the need for, nature of, and success of, conservation efforts. So insofar as conservation as a field is moving away from ecology as a field, maybe it’s not due to increasing specialization or “siloing”, but because conservation is becoming more interdisciplinary?

The Ohio state house has passed legislation prohibiting students from being “penalized or rewarded” on school assignments based on the religious content of their work. Local news is reporting this as allowing students to give scientifically-incorrect answers about evolution without losing marks, so long as their reasoning is religious. But it’s not clear to me if that’s the case, since the legislation also says that assignments “shall” be marked and grades calculated “using ordinary academic standards of substance and relevance, including any legitimate pedagogical concerns.” Anyone know more about this? Is this supposed to be a dog whistle to conservative Christians? An attempt to discourage teaching of evolution by creating a legal quagmire teachers won’t want to wade into? Or what?

The Ecology For The Masses blog is looking for new grad student and early career contributors.

This looks super-cool: Alexis Lerner’s stats course for Jewish Studies students. All about how quantitative information can complement, rather than oppose, other ways of knowing, such as firsthand personal experience. This is something I think a lot about, and wish I was better at conveying. (ht Andrew Gelman)

It turns out you can calculate the eigenvectors of (many) matrices directly from the eigenvalues. Holy crap. (EDIT: to elaborate a bit: the matrices for which it’s possible to calculate the eigenvectors from the eigenvalues are Hermitian matrices. Hermitian matrices crop up in many areas of science, hence my parenthetical “many”. And it also sounds like the calculation makes use of the eigenvalues of a minor matrix of the original matrix, not just the eigenvalues of the original matrix itself.)

I just found out about Simonton’s (2012) book Creativity In Science. Anyone read it? Should I read it?

Comedy Wildlife Photography Awards winner and finalists. These are hilarious! 🙂

DO IT. 🙂

7 thoughts on “Friday links: statistics vs. the humanities, ecology vs. conservation, and more

  1. I must be missing something on the claim that eigenvectors can be computed from eigenvalues — from my (very limited) understanding there is no mapping from one to the other. I could come at this many ways but here is one using PCA. We have an elipsoid-shaped scatter of points in 3D space (3 variables — imagine the points form a football in space). The PCs of this scatter are the eigenvectors and the variance on the points projected onto the PCs are the eigenvalues. If I rigidly rotate the ellipsoid in the 3D space the eigenvectors change but the eigenvalues don’t, so there is an infinite number of mappings from eigenvectors to the same eigenvalues. So how can we take the eigenvalues and get “the” eigenvectors? I need to read the article more closely but doing so makes my head spin (like the 3d scatter of points).

      • And the calculation also seems to involve the eigenvalues of a minor matrix of the original matrix, not just the eigenvalues of the original matrix itself. I have edited the post to clarify this.

      • Had to google hermitian matrix…the eigenvectors of a complex matrix also makes my head spin but at least I (think I) am correct that my logic above holds (because the result doesn’t apply to the matrices in my example)!

      • Yes, my understanding is that their result does not hold for the sorts of matrices you were thinking about.

        In ecology, Hermitian matrices actually don’t crop up often. I only happen to know about them because years ago I had a weird idea for an ecological application of them (that didn’t pan out).

  2. great comments on the HIntzen paper. To clarify, my main disagreement with this paper is that their methods don’t suit their aims – they are trying to identify how the relationship between disciplines has changed over time, & how ‘relevant’ ecology is to cons biol. You can’t do this with a content analysis of a bunch of subjectively-chosen journals. As a study showing how each discipline may have changed over time, sure. As for methods, agree with you, it is almost impossible to select ‘representative’ journals. I also wondered about focusing on individual researchers, i.e. identifying key ecologists & conservation biologists & analysing their publications over time, but this is just as subjective.

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 )

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.