6334 S 2014 Mayo/Spanos:
Short essay ~7 double-spaced pages
(DUE: March 26-April 3)
Choose a topic that combines some aspects of at least two of the readings (as of today or 3/27). You might connect two or three of the themes from the reading question sheets (with at least one being different from those you’ve turned in), or introduce new questions. You will have a chance to revise your paper. Ideally, it will form part of a final paper for this class, but you are also free to find something entirely new for the final paper.
This essay requires showing you understand and can synthesize some of the central philosophical issues we have discussed. Since it is a philosophical essay, attention should be paid to clarifying key terms. Use direct quotes from the readings. You can raise problems and questions, give suggestions as to where purported solutions seem to leave gaps (which you might take up later), but you need not, for this essay. You can mix statistical examples and discussion to the degree that you wish. I recommend you send me your tentative ideas or outline in the next few days, or come to my Wednesday office hours.
Some general topics:
- Problem of induction, inductive logics
- Popper, demarcation, selection effects (or any 2)
- Random variables and parameters, events and hypotheses
- Behavioristic vs inferential construal of tests and one or more criticisms of tests (e.g., their interpretation, fallacies of rejection, acceptance)
- Fisher/Neyman/Pearson-aspects of the historical debates and interpretation of confidence intervals or tests
- Some aspect of Bayesian vs non-Bayesian, highly probable vs highly probed
- Pre-data specifications and post-data interpretations
- Falsificationism: Fisher and Neyman, Fisher and Neyman-Pearson,
- Reinterpretating Popper: Falsifiability vs avoiding questionable science (avoiding tests that too-readily interpret accordance as evidence for a hypothesis)
- Reinterpreting tests (Mayo-Spanos) and ___
- P-values, the Higgs discovery and ___
- Criticisms and responses: Fisher vs. Neyman-Pearson, Meehl vs Popper, Howson and Urbach’s vs Mayo and Spanos, or other combinations
- Criticisms from HU (chapter 5): (e.g., confidence intervals use irrelevant information, HU 2006 175) well-supported hypothesis rejected in a significance test (HU 2006 154)
- Selection effects, tests of assumptions, Fisherian vs N-P tests
- Philosophy of probabilism and fraudbusting
Error Statistics Philosophy 