Anirban DasGupta has So. Many. Questions. He invites your answers: send your thoughts and comments (and maybe more questions!) to bulletin@imstat.org or leave a comment below.

Like a slightly impetuous, though scrupulous, young man of the teenage revolutionary kind, I am listing today an abundance of questions that more or less just occurred to me. I wish I could make at least a small fraction of IMS Bulletin readers spend a moment thinking about a few of them, if these questions have a nanoscopic grain of seriousness. 

I do not have probable answers to most of them. Maybe you do…

1. Should PhD students in a statistics department be required to know any mathematics beyond high school algebra, two semesters of calculus, and one semester of matrix calculations? If yes, what?

2. Should statistics departments give a common degree of PhD in statistics, or separate degrees of PhD in perhaps applied statistics, data science, biostatistics, algorithms and graphics?

3. Should there be a qualifying exam for PhD students in statistics? If yes, should it be a common qualifying exam for all students, or separate exams for separate areas?

4. If there are qualifying exams, how many chances should a student get to pass it? One, two, three? Infinite?

5. Should there be some external evaluation of PhD dissertations in statistics, such as reports by anonymous experts chosen by the university?

6. Should graduate school include two semesters of ethics and integrity classes?

7. Should we teach PhD students any finite sample optimality theory for given parametric models? Or, abandon it? How soon should they learn asymptotics? Should they be required to learn Bayes theory?

8. Alternatively, should we instead start our theory classes with model-free omnibus methods, such as the bootstrap?

9. Should students see proofs? Which proofs?

10. Should PhD students in statistics be required to know more probability than probability at the Hogg and Craig level? If yes, should we mandate theory of stationary Markov chains, random walks, martingales, Brownian motion, renewals, counting processes, diffusions, large deviations, Itô theory? Which ones?

11. Should PhD students in statistics be mandated to take classes in the CS department? In the math department? In the engineering school?

12. As a policy, should hiring committees be appointed, elected, or selected by a public closed-box random selection?

13. Should promotion committees be appointed, elected, or selected by a public randomization?

14. Should admission committees be appointed, elected, or selected by public randomization?

15. Should personnel committees be appointed, elected, or selected by public randomization?

16. Should hiring decisions be made by executives, or Qualtrics surveys, or a public vote counted by a set of faculty members?

17. Generally, should tenures of department heads be limited to at most three or four years, with no renewals?

18. Should departments maintain a physical room with recent publications of its faculty on display in shelves and racks and all receptions held in that room?

19. Should annual raises pass a test of uniformity with a p-value of .10 or so?

20. How about Gallup-like tracking polls of the approval rating of department heads among faculty members and PhD students?

21. Should department heads usually teach?

22. Should citations be considered at all? If yes, how? Total number? Average? Top five?

23. Should there be a list of statistics journals in which a faculty member must publish as a rule to get tenured, or promoted?

24. Should faculty members who are editors be automatically given teaching reduction?

25. And, finally, should faculty members be required to make good jokes?

Since 25 is the only multiple of five below 100 that is also a perfect square, this seems to be a good place to stop. Now this was astuteness in clean lines.