Also this week: how to increase graduation rates of students in financial need, Plos One’s surprisingly (?) high rejection rate, and more.
The fascinating history of research on “canals” on Mars–and how it began with a translation error.
Andrew Hendry on his most over-cited papers. He’s spot-on re: “fill in the box” citation inflation.
No detectable association between whether a randomized controlled trial was preregistered and whether it detected a positive treatment effect. (ht Retraction Watch)
An argument that you shouldn’t use funnel plots to diagnose publication bias, because funnel plots assume that study sample sizes are uncorrelated with the (true) effect size being studied. To which, yes, that’s an assumption of funnel plots. But unlike the authors of the linked post I think that assumption usually is fine. In my experience sample sizes usually are determined by considerations independent of the true effect size (in particular, the investigator’s resources). (ht Stephen Heard, via Twitter)
Wait, Plos One has a 50% rejection rate?! Even though they only reject on grounds of technical unsoundness? Anyone else surprised by this? Man, I sure hope that most other journals receive a lower rate of technically-unsound submissions than that (and I suspect they do…) Because the thought that 50% of all submissions are so technically flawed as to be unpublishable is pretty depressing.
Sorry for the delayed response, which includes various reasons for not replying, including “Sorry for the delay! I put off answering your e-mail until I had an even more tedious task that I wanted to avoid. Thanks!”
The NYTimes had a piece on a potential way to increase student graduation rates: giving small grants to needy students who are close to graduation. The piece focuses on Georgia State, which has done an impressive job of increasing graduation rates and erasing differences in graduation rates between white and minority students. The small grant program at Georgia State sounds really interesting, and I’m glad there are studies underway to see how much of an impact the small grants are having.