Also this week: Harrison Dyar’s wild life, gender biased credit for collaboration in some fields but not others, are all R users about to become Bayesians, unemployment vs. science, peak Plos One, and more.
Plos One is still huge–but its output is down 11% in the last two years. (ht Retraction Watch)
The rstanarm package, now on CRAN, implements Bayesian regularization approaches to some of the world’s most widely-used statistical methods (e.g., glm), using the same syntax as familiar R commands like lm(). So if WinBUGS was like giving neighborhood teenagers the key to your Ferrari (as Jim Clark or someone supposedly said), what is rstanarm? Giving neighborhood teenagers the keys to your Ferrari, liquor cabinet, and house? Giving neighborhood teenagers bus tickets? Or what? 🙂 We may be about to test Jeff Leek’s hypothesis that no statistical method works at scale. Cheap, unhelpful jokes aside, I actually have no useful opinion on this, so I’m keen to hear from those who do. In particular, if you want to know WWBBD, see here.
The idea of “overlaying” conventional peer-reviewed journals on arXiv in order to drive down publication costs is starting to take off. I’ve linked to news on this before. I remain curious as to how some (not all) open access advocates will react to this innovation, because it gives them something they want–cheap open access publishing–but does so by reinforcing something they’d like to get rid of (review that evaluates mss on criteria like “importance” and “novelty”). Also will be interesting to see whether such ventures take off, since they have to compete with established selective journals for good papers. (ht Retraction Watch)
Interesting piece from the Chronicle on how blogs have changed how academics communicate political science. Focuses on The Monkey Cage, which went from a mostly-academic blog to getting picked up by the Washington Post and aiming its output more at politicians, policymakers, and politically-engaged members of the general public. The piece also touches on how there’s been a shake-out from the burst of academic interest in blogging 10-15 years ago.
Wow: women in economics publish just as much as men–but they’re much more likely to be denied tenure at leading US universities. The difference persists if you control for differences in tenure rates across universities, the uneven distribution of women across subfields of economics, venue of publications, year of tenure decision, and other factors. Various lines of statistical evidence strongly suggest an explanation: women get much less credit from tenure committees for collaborative papers than men do, especially if the women co-author with men. In contrast, women and men in sociology get the same credit for collaborative work and are tenured at similar rates. And I’m stunned that one possible explanation is a simple difference in authorship practice: sociologists list authors in order of their contribution, while economists list authors alphabetically. Note that one caveat here is that tenure denials were inferred from from the relative ranks of the universities people moved to and from, for people who changed jobs 6-8 years after being hired. I worry that the results could be biased by gender differences in pre-tenure job switching (e.g., because of spousal issues). Which if so arguably just relocates the nature or effects of gender bias, but that matters–correctly diagnosing the problem is important because different diagnoses imply different solutions. Further discussion of this and some other interpretive details here. Caveat aside, I suspect that life science fields are more like sociology than economics, but it would be interesting to see data on that. (UPDATE: in most scientific fields, women who reach the asst. prof stage are slightly more likely than men to achieve tenure; see Shaw & Stanton 2012).
Better news on that front: success rates for men and women are similar at NSF and NIH. As long as you control for discipline, where scientists work, their academic rank, and years of experience. If you just look at raw numbers, women have lower success rates than men at NSF, but higher success rates than men at NIH.
Amy Parachnowitsch on how the first week of life as an unemployed academic wasn’t too different from life as an employed academic. I’ve been in Amy’s shoes, albeit only briefly. Hope it works out for you too, Amy.
Earlier this month I plugged Bob Trivers’ memoir of his wild life. I’d never heard of late 19th-century entomologist Harrison Dyar before, but now I want to read the new biography charting his wild life. He pioneered the reconstruction of insect evolutionary trees based on Darwinian principles–and was a bigamist who fought extensively with other entomologists.
Terry McGlynn with a great post on why to “network” with people in your field, and how you’re doing it wrong if you’re setting out with the goal of meeting famous people you feel like you “need” to know. One thing he doesn’t note is that just because networking matters does not mean that academic science is just some nepotistic game in which who you know and the science you do are two totally separate things, with the former trumping the latter. Echoes my own thoughts. I also have some old advice on how to network at conferences.
A rare retraction in ecology: a dispute over data use permission and other issues has led to the retraction of a paper estimating the abundance of a rare bird in Texas, over the objections of the paper’s authors. I have no comments, I find it difficult to judge such disputed situations as an outsider, based only on the necessarily-limited information in news reports.
Didham et al. 2015 cites one of Brian’s blog posts. (ht Allison Barner, via Twitter) Cool. It’s not the first time we’ve been cited; Peter Abrams recently framed a whole paper around one of Jeremy’s posts. Via Mark Vellend, some info on the (evolving) standards for how to cite blog posts here and here.
There’s little relationship between “believing in” human evolution and the ability to comprehend, be engaged by, and even convinced by, evolutionary science.
And finally, Notwitter. The site where you don’t publish all the things you shouldn’t say. (ht @dsquareddigest)
I found this piece on growth mindsets interesting. Encouraging a growth mindset is one key recommendation for countering stereotype threat, but people often are confused about what exactly a growth mindset is and how to encourage it. One section that stood out to me:
students often haven’t learned that working hard involves thinking hard, which involves reflecting on and changing our strategies so we become more and more effective learners over time, and we need to guide them to come to understand this. For example, a novice teacher who sees a student trying very hard but not making any progress may think “well, at least she’s working hard, so I’ll praise her effort,” but if the student continues to do what she’s doing, or even more of it, it’s unlikely to lead to success.
It also includes a link at the end to another interesting, recent piece on how understanding different types of mistakes can help us learn. And, on a similar topic, here’s a depressing tweet related to a new preprint:
I haven’t read the whole paper yet, and will be interested in hearing reactions from people who work on this sort of question.
This week, Congresswoman Jackie Speier gave a speech on the US House of Representatives floor on sexual harassment in science. Here’s an important follow up tweet: