Also this week: the best stats paper you’ll read this week was written in 1962, among-study heterogeneity vs. replicability, living is a way of life in Canada, the intuition behind carbon tax rebates, and more.
Roger Peng rereads Tukey’s (1962) “The future of data analysis”. Very interesting. He’s read it 17 times and continues to get new insights out of it, which has convinced me I need to read it at least once. Related to Brian’s old post on exploratory statistics. And to Meghan’s old post on which papers to read over and over.
High among-study heterogeneity doesn’t mean that the studies won’t replicate. Put another way, “generalizability” and “replicability” are two different concepts. I’d have thought that was obvious, but apparently not?
The reason zombie ideas (or really, all prominent ideas) persist is because so frickin’ many papers are published these days. At least, that’s the claim of a new unreviewed (and short) preprint: increasing volume of papers causes scholarly fields to ossify. I’ve skimmed it, the argument is very interesting. Basically, when you have too many papers being published, two things happen. First, authors come under increasing pressure to relate their work to previous work in order to have any hope of attracting even a scrap of attention from busy readers. Second, “clonal interference” among the flood of new ideas sets a speed limit to how fast the field as a whole can “evolve” (that’s not the author’s phrasing, that’s my own gloss.) The linked preprint analyzes masses of citation data that appear consistent with this hypothesis, though I don’t think they clinch the case. Thoughts?
The problem with trying to get people to care more about environmental causes by complaining about how much attention those causes get relative to how much the Notre Dame fire got (example). I agree 100% with this, kudos to Sam Perrin for being brave enough to say so, and to Kirsty MacLeod for first saying the same on Twitter (and maybe others have said the same and I didn’t see them; sorry to anyone I overlooked.) I’d broaden the point and argue that whataboutism (“You care about X. But what about Y?”) is not an effective way to get people to care more about your pet cause.
The intuition behind carbon taxes that are rebated to consumers. How can a tax that’s rebated create an incentive to change behavior? Even I–a PhD-holding ecologist who reads economics blogs–didn’t really “get” it until I read this post. Nick Rowe is a great explainer.
Women grant applicants to the Gates Foundation received lower scores than men on average even though the reviewers were blinded to proposal author identity (link goes to unreviewed preprint). This gender bias remains if you control for reviewer characteristics, proposal topic, or ex ante measures of applicant “quality”. But I don’t buy the authors’ proposed explanation (gender differences in word choice); the analysis is unconvincing to me for a bunch of reasons. So the only lesson I took away is, don’t just implement blind review and then assume that you’ll get gender-neutral outcomes. Gender-biased review outcomes can arise for lots of reasons besides reviewer knowledge of author gender, and if they do then blinding won’t produce gender-neutral outcomes (which isn’t an argument against blinded review, of course). Note that I wouldn’t assume that any of these results generalize to other funding agencies (or that they don’t). The data I’ve seen indicate that grant review outcomes at many (not all) funding agencies are gender-neutral, but I’m not an expert and haven’t exhaustively reviewed the literature. (ht Marginal Revolution).
This is from 2017, but I was just pointed to it and it’s relevant to stuff we’ve talked about recently, so I’m going to link to it. Against Purity by Alexis Shotwell (link goes to an interview with the author). It’s now on my reading list. Money quote:
[P]ersonal purity is simultaneously inadequate, impossible, and politically dangerous for shared projects of living on earth.
This week in Dumb Things I Refuse To Link To: none of y’all fell for that viral tweet falsely claiming that some non-existent prof was charging students money for recommendation letters, right?
This is very old, but honestly I probably do more good by calling attention to old stuff that’s still worth reading than I do by calling attention to stuff that many of you probably knew about already. So here’s an extended outtake and discussion of the funny, brilliant “In the New Canada, Living Is a Way of Life”. 🙂