Also this week: study preregistration vs. effect size, the Covid-19 questions we’re not asking, statistics lawn sign, and more.
Very sad news: evolutionary ecologist Josh Van Buskirk has passed away after a long battle with cancer. He was 62. This news hits me hard because Josh was one of my oldest friends in ecology. We shared an undergrad adviser, David Smith. Josh continued to collaborate with David for many years after graduating from Williams College, joining David every summer to conduct research on Isle Royale. I met Josh during my own summer on Isle Royale in 1994. Josh was a great guy–he was kind, personable, and smart. He was a keen outdoorsman–built his own canoes and kayaks. He was beyond a keen birder. He and David basically had the Audubon Guide to North American birds memorized (seriously). Josh did a lot of really good work in evolutionary ecology–he was an excellent experimenter. Among his most important work was pioneering research with David documenting adaptive plasticity in frog tadpole morphology and behavior. I was fortunate enough to get to watch some of that research as it was being done in the summer of 1994. That paper kicked off an entire research program taken up by many others, using anurans as a model system to study phenotypic plasticity and trait-mediated indirect effects. My path only crossed Josh’s occasionally in later years, but I always looked forward to catching up with him. Rest in peace Josh, I’ll miss you.
Biodiversity and Conservation has issued an Expression of Concern for Thompson & Newmaster 2014. This is the paper that first author Ken Thompson has been trying to get retracted, after discovering serious unexplained anomalies in the data provided by second author (and Thompson’s former undergrad adviser) Steven Newmaster. Here’s Ken’s blog post about the anomalies. Here’s Science’s news article on the case. And here’s our previous linkfest entry on the case.
I’m late to this, but here’s Musa al-Gharbi reviewing Stuart Ritchie’s recent book Science Fictions: How Fraud, Bias, Negligence, and Hype Undermine the Search for Truth. I need to put Ritchie’s book on my reading list.
Another case of prominent psychologist Dan Ariely claiming in interviews to have done studies that he not only never published, but seems never to have done. To which, I dunno man. It’s not as bad as publishing obviously-fake data. And I’m sure celebrities fib all the time in interviews, embellishing or even inventing anecdotes. But it creeps me out to see a scientist doing this.
The hard Covid-19 questions we’re not asking.
In psychology, effect sizes, measured as correlation coefficients, are less than half as large on average in pre-registered studies as in non-pre-registered studies. Gonna add this one to my list of statistical vignettes.
Jeremy, I only met Josh once on a trip to Zurich and everything you wrote about him rings true. I remember chatting about data analysis and Josh saying that he never did any analyses that he hadn’t explicitly stated he was going to do before he had collected the data. He felt that any ‘opportunistic’ analysis was a form of p-hacking. He struck me as a guy who was approaching his science with a great deal of integrity but without being self-righteous or self-congratulatory. It’s a loss for a ecology but a much greater one for his friends and family.
Yes, Josh was indeed a stickler for statistical rigor (in the best way–I mean that as a compliment).