André Ariew writes:

I’m a philosopher of science at the University of Missouri. I’m interested in leading a seminar on a variety of current topics with philosophical value, including problems with significance tests, the replication crisis, causation, correlation, randomized trials, etc. I’m hoping that you can point me in a good direction for accessible readings for the syllabus. Can you? While the course is at the graduate level, I don’t assume that my students are expert in the philosophy of science and likely don’t know what a p-value is (that’s the trouble—need to get people to understand these things). When I teach a course on inductive reasoning I typically assign Ian Hacking’s An Introduction to Probability and Inductive Logic. I’m familiar with the book and he’s a great historian and philosopher of science.

He’d like to do more:

Anything you might suggest would be greatly appreciated. I’ve always thought that issues like these are much more important to the philosophy of science than much of what passes as the standard corpus.

My response:

I’d start with the classic and very readable 2011 article by Simmons, Nelson, and Simonsohn, False-Positive Psychology: Undisclosed Flexibility in Data Collection and Analysis Allows Presenting Anything as Significant.

And follow up with my (subjective) historical overview from 2016, What has happened down here is the winds have changed.

You’ll want to assign at least one paper by Paul Meehl; here’s a link to a 1985 paper, and here’s a pointer to a paper from 1967, along with the question, “What happened? Why did it take us nearly 50 years to what Meehl was saying all along? This is what I want the intellectual history to help me understand,” and 137 comments in the discussion thread.

And I’ll also recommend my own three articles on the philosophy of statistics:

- [2017] Beyond subjective and objective in statistics (with discussion). {\em Journal of the Royal Statistical Society}. (Andrew Gelman and Christian Hennig)
- [2013] Philosophy and the practice of Bayesian statistics (with discussion). {\em British Journal of Mathematical and Statistical Psychology} {\bf 66}, 8–38. (Andrew Gelman and Cosma Shalizi)

[2013] Rejoinder to discussion. {\em British Journal of Mathematical and Statistical Psychology} {\bf 66}, 76–80. (Andrew Gelman and Cosma Shalizi) - [2011] Induction and deduction in Bayesian data analysis. {\em Rationality, Markets and Morals}, special topic issue “Statistical Science and Philosophy of Science: Where Do (Should) They Meet In 2011 and Beyond?”, ed.\ Deborah Mayo, Aris Spanos, and Kent Staley. (Andrew Gelman)

The last of these is the shortest so it might be a good place to start—or the only one, since it would be overkill to ask people to read all three.

Regarding p-values etc., the following article could be helpful (sorry, it’s another one of mine!):

- [2017] Some natural solutions to the p-value communication problem—and why they won’t work. {\em Journal of the American Statistical Association}. (Andrew Gelman and John Carlin)

And, for causation, I recommend these two articles, both of which should be readable for students without technical backgrounds:

- [2011] Experimental reasoning in social science. In {\em Field Experiments and their Critics}, ed.\ Dawn Teele. Yale University Press. (Andrew Gelman)
- [2011] Causality and statistical learning. {\em American Journal of Sociology}. (Andrew Gelman)

OK, that’ll get you started. Perhaps the commenters have further suggestions?

**P.S.** I’d love to lead a seminar on the philosophy of statistics, unfortunately I suspect that here at Columbia this would attract approximately 0 students. I do cover some of these issues in my class on Communicating Data and Statistics, though.

The post What readings should be included in a seminar on the philosophy of statistics, the replication crisis, causation, etc.? appeared first on Statistical Modeling, Causal Inference, and Social Science.