Repository URL:
Jan Sprenger
conference paper description
The impact of experimental design on the interpretation of a scientific experiment is a subject of major controversy. Are data a neutral arbiter between competing hypotheses, or is their interpretation intimately connected to the experimental design from which they are generated? The debate focuses on the relevance of stopping rules in sequential trials. However, Bayesian and frequentist statisticians and philosophers of science are apparently deadlocked in their controversy. To resolve the deadlock, I suggest a threefold strategy: (i) to distinguish various senses of relevance of stopping rules, (ii) to consider the requirements of experimental practice to a higher degree and (iii) to review the alleged counterexamples from a decision-theoretic perspective. While maintaining the pre-experimental relevance of design and stopping rules, this approach also leads us to the evidential, post-experimental irrelevance of stopping rules.

This conference paper has 0 Wikipedia mention.