I had occasion to refer to this post from a couple years ago on the anthropic principle in statistics. In that post, I wrote:
I actually used the anthropic principle in my 2000 article, Should we take measurements at an intermediate design point? (a paper that I love; but I just looked it up and it’s only been cited 3 times; that makes me so sad!), but without labeling the principle as such.
I looked it up, and . . . it still only has 3 citations! That’s some non-influential work we got there.
Procrastinating, I looked up the three. One of them spells your name wrong…
https://www.sciencedirect.com/science/article/pii/S0006322306010717
I’m guessing that the misspelling of Gelman as Gellman may have something to do with the fact that one of the authors is Barbara Geller!!
Wasn’t that one of the exam questions you posted one time?
https://statmodeling.stat.columbia.edu/2014/10/23/match-wits-columbia-stat-ph-d-students/
After 20 years have passed and the experimental design has often been rebranded into active learning, there are still many unknown questions even for this simple regression design. Classic experimental design literature is misleadingly focused on simple and discrete models (discrete factors, hence we can work with the true model). With continuous inputs, complex working models (e.g., a GP regression rather than a linear regression), an M-open view, and sequential data collection, it is infeasible to derive a closed-form optimal design by hand—yet we still need some guidance on experimental design in practice.