Syndicate content

Weekly links May 13: Bad science reporting, bad measurement, bad social science ethics, and good books

David Evans's picture
 
  • This week in macro measurement: “‘Laws are like sausages, it is better not to see them being made’ said Otto von Bismarck. Turns out you can probably add GDP to that list.” Duncan Green gives a useful summary of The Economist’s extensive critique of GDP, how it is becoming decreasingly useful over time, and how it could be better.
  • This week in bad social science: “A group of researchers has released a data set on nearly 70,000 users of the online dating site OkCupid. The data dump breaks the cardinal rule of social science research ethics: It took identifiable personal data without permission.” Brian Resnick at Vox has a useful rundown of a serious privacy breach. On the other hand, the master’s student at the center of the controversy edits a journal (Open Differential Psychology) where half the articles are by him. So that’s an idea.
   
  • This week in conflict: “It’s been 25 years since Sierra Leone’s brutal civil war. Here’s what we know about helping communities recover.” Rachel Glennerster has a useful write-up of work from Cilliers, Dube, & Siddiqi (out today in Science) on “community-based reconciliation efforts in Sierra Leone,” contrasting them with community-driven reconstruction efforts that she evaluated. “Rebuilding nations after war must attend to the repair not just of physical infrastructure but also of communities and individuals. We need innovative ways to help when thousands of people have experienced extraordinary levels of trauma.”
 
  • This week in literature: Glennerster’s piece reminds me of Aminatta Forna’s brilliant novel The Memory of Love, about pre- and post-war Sierra Leone, also recommended to me by Glennerster and reviewed by me here. At one point in the novel, international researchers deliver a report estimating that “ninety-nine per cent of the population was suffering from post-traumatic stress disorder.”
 
  • This week in socioemotional skills: Angela Duckworth’s book, Grit: The Power of Passion and Perseverance, is out this month. (The TED talk has been out for a while.) In a talk at the World Bank, Duckworth described going to some low-performing U.S. schools, where she observed no teaching happening, and commented –I obviously paraphrase – that these schools needed a lot of things in place before they should worry about Duckworth providing grit training. While the excitement about socio-emotional skills (of which family grit is a member) is warranted by evidence that these matter a lot (see Heckman et al. among many other places), Duckworth has cautioned in the New York Times against rushing to holding schools accountable for this, partly because we’re still figuring out how much we can change grit, and partly because of difficulty in measurement and reporting.
 
  • And finally, this week in bad reporting on science: “Science is constantly producing new studies, as you know if you ever watch TV.” John Oliver takes down media reporting on science in this 20-minute, highly entertaining video. I’m reminded of Jeff Mosenkis’s very useful “guide to debunking debunking news stories” (yes, debunking is supposed to be there twice). It’s Friday: If you haven’t watched this yet, treat yourself. He also touches on p-hacking. The best comic treatment of p-hacking comes from xkcd, on jelly beans and acne.
   

Add new comment